IPv4/6 reassembly

Change-Id: Ic5dcadd13c88b8a5e7896dab82404509c081614a
Signed-off-by: Klement Sekera <ksekera@cisco.com>

20 files changed, 2866 insertions, 8 deletions

diff --git a/src/vnet/buffer.h b/src/vnet/buffer.h
index c9f1d6ef6a1..5d7273aa04f 100644
--- a/src/vnet/buffer.h
+++ b/src/vnet/buffer.h
@@ -177,6 +177,18 @@ typedef struct
 	  u8 code;
 	  u32 data;
 	} icmp;
+
+	/* reassembly */
+	struct
+	{
+	  u16 fragment_first;
+	  u16 fragment_last;
+	  u16 range_first;
+	  u16 range_last;
+	  u32 next_range_bi;
+	  u16 ip6_frag_hdr_offset;
+	  u16 estimated_mtu;
+	} reass;
       };
 
     } ip;
diff --git a/src/vnet/ip/icmp46_packet.h b/src/vnet/ip/icmp46_packet.h
index a86cbd57bdb..e761af8bf16 100644
--- a/src/vnet/ip/icmp46_packet.h
+++ b/src/vnet/ip/icmp46_packet.h
@@ -149,6 +149,7 @@
   _ (parameter_problem, 0, erroneous_header_field)			\
   _ (parameter_problem, 1, unrecognized_next_header)			\
   _ (parameter_problem, 2, unrecognized_option)				\
+  _ (parameter_problem, 3, first_fragment_has_incomplete_header_chain)  \
   _ (router_renumbering, 0, command)					\
   _ (router_renumbering, 1, result)					\
   _ (node_information_request, 0, data_contains_ip6_address)		\
diff --git a/src/vnet/ip/icmp6.h b/src/vnet/ip/icmp6.h
index 9a3487b1c5c..1cac8240466 100644
--- a/src/vnet/ip/icmp6.h
+++ b/src/vnet/ip/icmp6.h
@@ -45,7 +45,7 @@
   _ (DEST_UNREACH_SENT, "destination unreachable response sent")	\
   _ (PACKET_TOO_BIG_SENT, "packet too big response sent")		\
   _ (TTL_EXPIRE_SENT, "hop limit exceeded response sent")		\
-  _ (PARAM_PROBLEM_SENT, "parameter Pproblem response sent")		\
+  _ (PARAM_PROBLEM_SENT, "parameter problem response sent")		\
   _ (DROP, "error message dropped")
 
 
diff --git a/src/vnet/ip/ip.api b/src/vnet/ip/ip.api
index 0da0a61fb75..6ed5a9d6bdb 100644
--- a/src/vnet/ip/ip.api
+++ b/src/vnet/ip/ip.api
@@ -812,6 +812,34 @@ autoreply define ioam_disable
   u16 id;
 };
 
+autoreply define ip_reassembly_set
+{
+  u32 client_index;
+  u32 context;
+  u32 timeout_ms;
+  u32 max_reassemblies;
+  u32 expire_walk_interval_ms;
+  u8 is_ip6;
+};
+
+define ip_reassembly_get
+{
+  u32 client_index;
+  u32 context;
+  u8 is_ip6;
+};
+
+define ip_reassembly_get_reply
+{
+  u32 client_index;
+  u32 context;
+  i32 retval;
+  u32 timeout_ms;
+  u32 max_reassemblies;
+  u32 expire_walk_interval_ms;
+  u8 is_ip6;
+};
+
 /*
  * Local Variables:
  * eval: (c-set-style "gnu")
diff --git a/src/vnet/ip/ip4_error.h b/src/vnet/ip/ip4_error.h
index 4e545aeda3c..00b1c6daad5 100644
--- a/src/vnet/ip/ip4_error.h
+++ b/src/vnet/ip/ip4_error.h
@@ -71,14 +71,19 @@
   _ (UNICAST_SOURCE_CHECK_FAILS, "ip4 unicast source check fails")	\
                                                                         \
   /* Spoofed packets in ip4-rewrite-local */                            \
-  _(SPOOFED_LOCAL_PACKETS, "ip4 spoofed local-address packet drops")    \
+  _ (SPOOFED_LOCAL_PACKETS, "ip4 spoofed local-address packet drops")   \
                                                                         \
   /* Errors singalled by ip4-inacl */                                   \
   _ (INACL_TABLE_MISS, "input ACL table-miss drops")                    \
   _ (INACL_SESSION_DENY, "input ACL session deny drops")                \
                                                                         \
   /* Erros from mfib-forward */                                         \
-  _ (RPF_FAILURE, "Multicast RPF check failed")
+  _ (RPF_FAILURE, "Multicast RPF check failed")                         \
+                                                                        \
+  /* Errors signalled by ip4-reassembly */                              \
+  _ (REASS_DUPLICATE_FRAGMENT, "duplicate/overlapping fragments")       \
+  _ (REASS_LIMIT_REACHED, "drops due to concurrent reassemblies limit") \
+  _ (REASS_TIMEOUT, "fragments dropped due to reassembly timeout")
 
 typedef enum
 {
diff --git a/src/vnet/ip/ip4_forward.c b/src/vnet/ip/ip4_forward.c
index 3ddf6dfa783..2a21135bb3e 100755
--- a/src/vnet/ip/ip4_forward.c
+++ b/src/vnet/ip/ip4_forward.c
@@ -1582,8 +1582,12 @@ ip4_local_inline (vlib_main_t * vm,
 
 	  /* Treat IP frag packets as "experimental" protocol for now
 	     until support of IP frag reassembly is implemented */
-	  proto0 = ip4_is_fragment (ip0) ? 0xfe : ip0->protocol;
-	  proto1 = ip4_is_fragment (ip1) ? 0xfe : ip1->protocol;
+	  proto0 =
+	    ip4_is_fragment (ip0) ? IP_PROTOCOL_VPP_FRAGMENTATION :
+	    ip0->protocol;
+	  proto1 =
+	    ip4_is_fragment (ip1) ? IP_PROTOCOL_VPP_FRAGMENTATION :
+	    ip1->protocol;
 
 	  if (head_of_feature_arc == 0)
 	    goto skip_checks;
@@ -1748,7 +1752,9 @@ ip4_local_inline (vlib_main_t * vm,
 
 	  /* Treat IP frag packets as "experimental" protocol for now
 	     until support of IP frag reassembly is implemented */
-	  proto0 = ip4_is_fragment (ip0) ? 0xfe : ip0->protocol;
+	  proto0 =
+	    ip4_is_fragment (ip0) ? IP_PROTOCOL_VPP_FRAGMENTATION :
+	    ip0->protocol;
 
 	  if (head_of_feature_arc == 0 || p0->flags & VNET_BUFFER_F_IS_NATED)
 	    goto skip_check;
@@ -1839,6 +1845,7 @@ VLIB_REGISTER_NODE (ip4_local_node) =
     [IP_LOCAL_NEXT_PUNT] = "ip4-punt",
     [IP_LOCAL_NEXT_UDP_LOOKUP] = "ip4-udp-lookup",
     [IP_LOCAL_NEXT_ICMP] = "ip4-icmp-input",
+    [IP_LOCAL_NEXT_REASSEMBLY] = "ip4-reassembly",
   },
 };
 /* *INDENT-ON* */
diff --git a/src/vnet/ip/ip4_input.c b/src/vnet/ip/ip4_input.c
index 121f40f473d..61c86ef112d 100644
--- a/src/vnet/ip/ip4_input.c
+++ b/src/vnet/ip/ip4_input.c
@@ -289,6 +289,7 @@ VLIB_REGISTER_NODE (ip4_input_node) = {
     [IP4_INPUT_NEXT_LOOKUP] = "ip4-lookup",
     [IP4_INPUT_NEXT_LOOKUP_MULTICAST] = "ip4-mfib-forward-lookup",
     [IP4_INPUT_NEXT_ICMP_ERROR] = "ip4-icmp-error",
+    [IP4_INPUT_NEXT_REASSEMBLY] = "ip4-reassembly",
   },
 
   .format_buffer = format_ip4_header,
@@ -311,6 +312,7 @@ VLIB_REGISTER_NODE (ip4_input_no_checksum_node,static) = {
     [IP4_INPUT_NEXT_LOOKUP] = "ip4-lookup",
     [IP4_INPUT_NEXT_LOOKUP_MULTICAST] = "ip4-mfib-forward-lookup",
     [IP4_INPUT_NEXT_ICMP_ERROR] = "ip4-icmp-error",
+    [IP4_INPUT_NEXT_REASSEMBLY] = "ip4-reassembly",
   },
 
   .format_buffer = format_ip4_header,
diff --git a/src/vnet/ip/ip4_input.h b/src/vnet/ip/ip4_input.h
index 75306a34db3..2692848f8ad 100644
--- a/src/vnet/ip/ip4_input.h
+++ b/src/vnet/ip/ip4_input.h
@@ -52,6 +52,7 @@ typedef enum
   IP4_INPUT_NEXT_LOOKUP,
   IP4_INPUT_NEXT_LOOKUP_MULTICAST,
   IP4_INPUT_NEXT_ICMP_ERROR,
+  IP4_INPUT_NEXT_REASSEMBLY,
   IP4_INPUT_N_NEXT,
 } ip4_input_next_t;
 
diff --git a/src/vnet/ip/ip4_reassembly.c b/src/vnet/ip/ip4_reassembly.c
new file mode 100644
index 00000000000..6b8d66532fa
--- /dev/null
+++ b/src/vnet/ip/ip4_reassembly.c
@@ -0,0 +1,1277 @@
+/*
+ * Copyright (c) 2017 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 IPv4 Reassembly.
+ *
+ * This file contains the source code for IPv4 reassembly.
+ */
+
+#include <vppinfra/vec.h>
+#include <vnet/vnet.h>
+#include <vnet/ip/ip.h>
+#include <vppinfra/bihash_24_8.h>
+#include <vnet/ip/ip4_reassembly.h>
+
+#define MSEC_PER_SEC 1000
+#define IP4_REASS_TIMEOUT_DEFAULT_MS 100
+#define IP4_REASS_EXPIRE_WALK_INTERVAL_DEFAULT_MS 10000	// 10 seconds default
+#define IP4_REASS_MAX_REASSEMBLIES_DEAFULT 1024
+#define IP4_REASS_HT_LOAD_FACTOR (0.75)
+
+#define IP4_REASS_DEBUG_BUFFERS 0
+#if IP4_REASS_DEBUG_BUFFERS
+#define IP4_REASS_DEBUG_BUFFER(bi, what)             \
+  do                                                 \
+    {                                                \
+      u32 _bi = bi;                                  \
+      printf (#what "buffer %u", _bi);               \
+      vlib_buffer_t *_b = vlib_get_buffer (vm, _bi); \
+      while (_b->flags & VLIB_BUFFER_NEXT_PRESENT)   \
+        {                                            \
+          _bi = _b->next_buffer;                     \
+          printf ("[%u]", _bi);                      \
+          _b = vlib_get_buffer (vm, _bi);            \
+        }                                            \
+      printf ("\n");                                 \
+      fflush (stdout);                               \
+    }                                                \
+  while (0)
+#else
+#define IP4_REASS_DEBUG_BUFFER(...)
+#endif
+
+static vlib_node_registration_t ip4_reass_node;
+
+typedef struct
+{
+  union
+  {
+    struct
+    {
+      // align by making this 4 octets even though its a 2 octets field
+      u32 xx_id;
+      ip4_address_t src;
+      ip4_address_t dst;
+      // align by making this 4 octets even though its a 2 octets field
+      u32 frag_id;
+      // align by making this 4 octets even though its a 1 octet field
+      u32 proto;
+      u32 unused;
+    };
+    u64 as_u64[3];
+  };
+} ip4_reass_key_t;
+
+always_inline u32
+ip4_reass_buffer_get_data_offset_no_check (vlib_buffer_t * b)
+{
+  vnet_buffer_opaque_t *vnb = vnet_buffer (b);
+  return vnb->ip.reass.range_first - vnb->ip.reass.fragment_first;
+}
+
+always_inline u32
+ip4_reass_buffer_get_data_offset (vlib_buffer_t * b)
+{
+  vnet_buffer_opaque_t *vnb = vnet_buffer (b);
+  ASSERT (vnb->ip.reass.range_first >= vnb->ip.reass.fragment_first);
+  return ip4_reass_buffer_get_data_offset_no_check (b);
+}
+
+always_inline u16
+ip4_reass_buffer_get_data_len_no_check (vlib_buffer_t * b)
+{
+  vnet_buffer_opaque_t *vnb = vnet_buffer (b);
+  return clib_min (vnb->ip.reass.range_last, vnb->ip.reass.fragment_last) -
+    (vnb->ip.reass.fragment_first + ip4_reass_buffer_get_data_offset (b)) + 1;
+}
+
+always_inline u16
+ip4_reass_buffer_get_data_len (vlib_buffer_t * b)
+{
+  vnet_buffer_opaque_t *vnb = vnet_buffer (b);
+  ASSERT (vnb->ip.reass.range_last > vnb->ip.reass.fragment_first);
+  return ip4_reass_buffer_get_data_len_no_check (b);
+}
+
+typedef struct
+{
+  // hash table key
+  ip4_reass_key_t key;
+  f64 first_heard;
+  // time when last packet was received
+  f64 last_heard;
+  // internal id of this reassembly
+  u32 id;
+  // buffer index of first buffer in this reassembly context
+  u32 first_bi;
+  // last octet of packet, ~0 until fragment without more_fragments arrives
+  u32 last_packet_octet;
+  // length of data collected so far
+  u32 data_len;
+  // trace operation counter
+  u32 trace_op_counter;
+} ip4_reass_t;
+
+typedef struct
+{
+  // IPv4 config
+  u32 timeout_ms;
+  f64 timeout;
+  u32 expire_walk_interval_ms;
+  u32 max_reass_n;
+
+  // IPv4 runtime
+  ip4_reass_t *pool;
+  clib_bihash_24_8_t hash;
+  u32 reass_n;
+  u32 id_counter;
+  u32 buffers_n;
+
+  // convenience
+  vlib_main_t *vlib_main;
+  vnet_main_t *vnet_main;
+
+  // node index of ip4-drop node
+  u32 ip4_drop_idx;
+  u32 ip4_reass_expire_node_idx;
+
+} ip4_reass_main_t;
+
+ip4_reass_main_t ip4_reass_main;
+
+typedef enum
+{
+  IP4_REASSEMBLY_NEXT_INPUT,
+  IP4_REASSEMBLY_NEXT_DROP,
+  IP4_REASSEMBLY_N_NEXT,
+} ip4_reass_next_t;
+
+typedef enum
+{
+  RANGE_NEW,
+  RANGE_SHRINK,
+  RANGE_DISCARD,
+  RANGE_OVERLAP,
+  FINALIZE,
+} ip4_reass_trace_operation_e;
+
+typedef struct
+{
+  u16 range_first;
+  u16 range_last;
+  u32 range_bi;
+  i32 data_offset;
+  u32 data_len;
+  u32 first_bi;
+} ip4_reass_range_trace_t;
+
+typedef struct
+{
+  ip4_reass_trace_operation_e action;
+  u32 pool_index;
+  u32 reass_id;
+  ip4_reass_range_trace_t trace_range;
+  u32 size_diff;
+  u32 op_id;
+  u32 fragment_first;
+  u32 fragment_last;
+  u32 total_data_len;
+} ip4_reass_trace_t;
+
+void
+ip4_reass_trace_details (vlib_main_t * vm, u32 bi,
+			 ip4_reass_range_trace_t * trace)
+{
+  vlib_buffer_t *b = vlib_get_buffer (vm, bi);
+  vnet_buffer_opaque_t *vnb = vnet_buffer (b);
+  trace->range_first = vnb->ip.reass.range_first;
+  trace->range_last = vnb->ip.reass.range_last;
+  trace->data_offset = ip4_reass_buffer_get_data_offset_no_check (b);
+  trace->data_len = ip4_reass_buffer_get_data_len_no_check (b);
+  trace->range_bi = bi;
+}
+
+u8 *
+format_ip4_reass_range_trace (u8 * s, va_list * args)
+{
+  ip4_reass_range_trace_t *trace = va_arg (*args, ip4_reass_range_trace_t *);
+  s = format (s, "range: [%u, %u], off %d, len %u, bi %u", trace->range_first,
+	      trace->range_last, trace->data_offset, trace->data_len,
+	      trace->range_bi);
+  return s;
+}
+
+u8 *
+format_ip4_reass_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 *);
+  ip4_reass_trace_t *t = va_arg (*args, ip4_reass_trace_t *);
+  s = format (s, "reass id: %u, op id: %u ", t->reass_id, t->op_id);
+  u32 indent = format_get_indent (s);
+  s = format (s, "first bi: %u, data len: %u, ip/fragment[%u, %u]",
+	      t->trace_range.first_bi, t->total_data_len, t->fragment_first,
+	      t->fragment_last);
+  switch (t->action)
+    {
+    case RANGE_SHRINK:
+      s = format (s, "\n%Ushrink %U by %u", format_white_space, indent,
+		  format_ip4_reass_range_trace, &t->trace_range,
+		  t->size_diff);
+      break;
+    case RANGE_DISCARD:
+      s = format (s, "\n%Udiscard %U", format_white_space, indent,
+		  format_ip4_reass_range_trace, &t->trace_range);
+      break;
+    case RANGE_NEW:
+      s = format (s, "\n%Unew %U", format_white_space, indent,
+		  format_ip4_reass_range_trace, &t->trace_range);
+      break;
+    case RANGE_OVERLAP:
+      s = format (s, "\n%Uoverlapping/ignored %U", format_white_space, indent,
+		  format_ip4_reass_range_trace, &t->trace_range);
+      break;
+    case FINALIZE:
+      s = format (s, "\n%Ufinalize reassembly", format_white_space, indent);
+      break;
+    }
+  return s;
+}
+
+void
+ip4_reass_add_trace (vlib_main_t * vm, vlib_node_runtime_t * node,
+		     ip4_reass_main_t * rm, ip4_reass_t * reass, u32 bi,
+		     ip4_reass_trace_operation_e action, u32 size_diff)
+{
+  vlib_buffer_t *b = vlib_get_buffer (vm, bi);
+  vnet_buffer_opaque_t *vnb = vnet_buffer (b);
+  ip4_reass_trace_t *t = vlib_add_trace (vm, node, b, sizeof (t[0]));
+  t->pool_index = reass - rm->pool;
+  t->reass_id = reass->id;
+  t->action = action;
+  ip4_reass_trace_details (vm, bi, &t->trace_range);
+  t->size_diff = size_diff;
+  t->op_id = reass->trace_op_counter;
+  ++reass->trace_op_counter;
+  t->fragment_first = vnb->ip.reass.fragment_first;
+  t->fragment_last = vnb->ip.reass.fragment_last;
+  t->trace_range.first_bi = reass->first_bi;
+  t->total_data_len = reass->data_len;
+#if 0
+  static u8 *s = NULL;
+  s = format (s, "%U", format_ip4_reass_trace, NULL, NULL, t);
+  printf ("%.*s\n", vec_len (s), s);
+  fflush (stdout);
+  vec_reset_length (s);
+#endif
+}
+
+void
+ip4_reass_free (ip4_reass_main_t * rm, ip4_reass_t * reass)
+{
+  clib_bihash_kv_24_8_t kv;
+  kv.key[0] = reass->key.as_u64[0];
+  kv.key[1] = reass->key.as_u64[1];
+  kv.key[2] = reass->key.as_u64[2];
+  clib_bihash_add_del_24_8 (&rm->hash, &kv, 0);
+  pool_put (rm->pool, reass);
+  --rm->reass_n;
+}
+
+static void
+ip4_reass_on_timeout (vlib_main_t * vm, ip4_reass_main_t * rm,
+		      ip4_reass_t * reass, u32 ** vec_drop_timeout)
+{
+  u32 range_bi = reass->first_bi;
+  vlib_buffer_t *range_b;
+  vnet_buffer_opaque_t *range_vnb;
+  while (~0 != range_bi)
+    {
+      range_b = vlib_get_buffer (vm, range_bi);
+      range_vnb = vnet_buffer (range_b);
+      u32 bi = range_bi;
+      while (~0 != bi)
+	{
+	  vec_add1 (*vec_drop_timeout, bi);
+	  vlib_buffer_t *b = vlib_get_buffer (vm, bi);
+	  if (b->flags & VLIB_BUFFER_NEXT_PRESENT)
+	    {
+	      bi = b->next_buffer;
+	      b->flags &= ~VLIB_BUFFER_NEXT_PRESENT;
+	    }
+	  else
+	    {
+	      bi = ~0;
+	    }
+	}
+      range_bi = range_vnb->ip.reass.next_range_bi;
+    }
+}
+
+ip4_reass_t *
+ip4_reass_find_or_create (vlib_main_t * vm, ip4_reass_main_t * rm,
+			  ip4_reass_key_t * k, u32 ** vec_drop_timeout)
+{
+  ip4_reass_t *reass = NULL;
+  f64 now = vlib_time_now (rm->vlib_main);
+  clib_bihash_kv_24_8_t kv, value;
+  kv.key[0] = k->as_u64[0];
+  kv.key[1] = k->as_u64[1];
+  kv.key[2] = k->as_u64[2];
+
+  if (!clib_bihash_search_24_8 (&rm->hash, &kv, &value))
+    {
+      reass = pool_elt_at_index (rm->pool, value.value);
+      if (now > reass->last_heard + rm->timeout)
+	{
+	  ip4_reass_on_timeout (vm, rm, reass, vec_drop_timeout);
+	  ip4_reass_free (rm, reass);
+	  reass = NULL;
+	}
+    }
+
+  if (reass)
+    {
+      reass->last_heard = now;
+      return reass;
+    }
+
+  if (rm->reass_n >= rm->max_reass_n)
+    {
+      reass = NULL;
+      return reass;
+    }
+  else
+    {
+      pool_get (rm->pool, reass);
+      memset (reass, 0, sizeof (*reass));
+      reass->id = rm->id_counter;
+      ++rm->id_counter;
+      reass->first_bi = ~0;
+      reass->last_packet_octet = ~0;
+      reass->data_len = 0;
+      ++rm->reass_n;
+    }
+
+  reass->key.as_u64[0] = kv.key[0] = k->as_u64[0];
+  reass->key.as_u64[1] = kv.key[1] = k->as_u64[1];
+  reass->key.as_u64[2] = kv.key[2] = k->as_u64[2];
+  kv.value = reass - rm->pool;
+  reass->last_heard = now;
+
+  if (clib_bihash_add_del_24_8 (&rm->hash, &kv, 1))
+    {
+      ip4_reass_free (rm, reass);
+      reass = NULL;
+    }
+
+  return reass;
+}
+
+void
+ip4_reass_finalize (vlib_main_t * vm, vlib_node_runtime_t * node,
+		    ip4_reass_main_t * rm, ip4_reass_t * reass, u32 * bi0,
+		    u32 * next0, vlib_error_t * error0, u32 next_input,
+		    u32 ** vec_drop_compress, u32 ** vec_drop_overlap)
+{
+  ASSERT (~0 != reass->first_bi);
+  vlib_buffer_t *first_b = vlib_get_buffer (vm, reass->first_bi);
+  vlib_buffer_t *last_b = NULL;
+  u32 sub_chain_bi = reass->first_bi;
+  u32 total_length = 0;
+  u32 buf_cnt = 0;
+  u32 dropped_cnt = 0;
+  do
+    {
+      u32 tmp_bi = sub_chain_bi;
+      vlib_buffer_t *tmp = vlib_get_buffer (vm, tmp_bi);
+      ip4_header_t *ip = vlib_buffer_get_current (tmp);
+      u32 data_len = ip4_reass_buffer_get_data_len (tmp);
+      u32 trim_front =
+	ip4_header_bytes (ip) + ip4_reass_buffer_get_data_offset (tmp);
+      u32 trim_end =
+	vlib_buffer_length_in_chain (vm, tmp) - trim_front - data_len;
+      if (tmp_bi == reass->first_bi)
+	{
+	  /* first buffer - keep ip4 header */
+	  ASSERT (0 == ip4_reass_buffer_get_data_offset (tmp));
+	  trim_front = 0;
+	  trim_end = vlib_buffer_length_in_chain (vm, tmp) - data_len -
+	    ip4_header_bytes (ip);
+	}
+      u32 keep_data =
+	vlib_buffer_length_in_chain (vm, tmp) - trim_front - trim_end;
+      while (1)
+	{
+	  ++buf_cnt;
+	  if (trim_front)
+	    {
+	      if (trim_front > tmp->current_length)
+		{
+		  /* drop whole buffer */
+		  vec_add1 (*vec_drop_compress, tmp_bi);
+		  ++dropped_cnt;
+		  trim_front -= tmp->current_length;
+		  ASSERT (tmp->flags & VLIB_BUFFER_NEXT_PRESENT);
+		  tmp->flags &= ~VLIB_BUFFER_NEXT_PRESENT;
+		  tmp_bi = tmp->next_buffer;
+		  tmp = vlib_get_buffer (vm, tmp_bi);
+		  continue;
+		}
+	      else
+		{
+		  vlib_buffer_advance (tmp, trim_front);
+		  trim_front = 0;
+		}
+	    }
+	  if (keep_data)
+	    {
+	      if (last_b)
+		{
+		  last_b->flags |= VLIB_BUFFER_NEXT_PRESENT;
+		  last_b->next_buffer = tmp_bi;
+		}
+	      last_b = tmp;
+	      if (keep_data <= tmp->current_length)
+		{
+		  tmp->current_length = keep_data;
+		  keep_data = 0;
+		}
+	      else
+		{
+		  keep_data -= tmp->current_length;
+		  ASSERT (tmp->flags & VLIB_BUFFER_NEXT_PRESENT);
+		}
+	      total_length += tmp->current_length;
+	    }
+	  else
+	    {
+	      vec_add1 (*vec_drop_overlap, tmp_bi);
+	      ASSERT (reass->first_bi != tmp_bi);
+	      ++dropped_cnt;
+	    }
+	  if (tmp->flags & VLIB_BUFFER_NEXT_PRESENT)
+	    {
+	      tmp_bi = tmp->next_buffer;
+	      tmp = vlib_get_buffer (vm, tmp->next_buffer);
+	    }
+	  else
+	    {
+	      break;
+	    }
+	}
+      sub_chain_bi =
+	vnet_buffer (vlib_get_buffer (vm, sub_chain_bi))->ip.
+	reass.next_range_bi;
+    }
+  while (~0 != sub_chain_bi);
+  last_b->flags &= ~VLIB_BUFFER_NEXT_PRESENT;
+  ASSERT (rm->buffers_n >= (buf_cnt - dropped_cnt));
+  rm->buffers_n -= buf_cnt - dropped_cnt;
+  ASSERT (total_length >= first_b->current_length);
+  total_length -= first_b->current_length;
+  first_b->flags |= VLIB_BUFFER_TOTAL_LENGTH_VALID;
+  first_b->total_length_not_including_first_buffer = total_length;
+  ip4_header_t *ip = vlib_buffer_get_current (first_b);
+  ip->flags_and_fragment_offset = 0;
+  ip->length = clib_host_to_net_u16 (first_b->current_length + total_length);
+  ip->checksum = ip4_header_checksum (ip);
+  vlib_buffer_chain_compress (vm, first_b, vec_drop_compress);
+  if (PREDICT_FALSE (first_b->flags & VLIB_BUFFER_IS_TRACED))
+    {
+      ip4_reass_add_trace (vm, node, rm, reass, reass->first_bi, FINALIZE, 0);
+#if 0
+      // following code does a hexdump of packet fragments to stdout ...
+      do
+	{
+	  u32 bi = reass->first_bi;
+	  u8 *s = NULL;
+	  while (~0 != bi)
+	    {
+	      vlib_buffer_t *b = vlib_get_buffer (vm, bi);
+	      s = format (s, "%u: %U\n", bi, format_hexdump,
+			  vlib_buffer_get_current (b), b->current_length);
+	      if (b->flags & VLIB_BUFFER_NEXT_PRESENT)
+		{
+		  bi = b->next_buffer;
+		}
+	      else
+		{
+		  break;
+		}
+	    }
+	  printf ("%.*s\n", vec_len (s), s);
+	  fflush (stdout);
+	  vec_free (s);
+	}
+      while (0);
+#endif
+    }
+  *bi0 = reass->first_bi;
+  *next0 = next_input;
+  *error0 = IP4_ERROR_NONE;
+  ip4_reass_free (rm, reass);
+  reass = NULL;
+}
+
+static u32
+ip4_reass_get_buffer_chain_length (vlib_main_t * vm, vlib_buffer_t * b)
+{
+  u32 len = 0;
+  while (b)
+    {
+      ++len;
+      if (PREDICT_FALSE (b->flags & VLIB_BUFFER_NEXT_PRESENT))
+	{
+	  b = vlib_get_buffer (vm, b->next_buffer);
+	}
+      else
+	{
+	  break;
+	}
+    }
+  return len;
+}
+
+static void
+ip4_reass_insert_range_in_chain (vlib_main_t * vm,
+				 ip4_reass_main_t * rm,
+				 ip4_reass_t * reass,
+				 u32 prev_range_bi, u32 new_next_bi)
+{
+
+  vlib_buffer_t *new_next_b = vlib_get_buffer (vm, new_next_bi);
+  vnet_buffer_opaque_t *new_next_vnb = vnet_buffer (new_next_b);
+  if (~0 != prev_range_bi)
+    {
+      vlib_buffer_t *prev_b = vlib_get_buffer (vm, prev_range_bi);
+      vnet_buffer_opaque_t *prev_vnb = vnet_buffer (prev_b);
+      new_next_vnb->ip.reass.next_range_bi = prev_vnb->ip.reass.next_range_bi;
+      prev_vnb->ip.reass.next_range_bi = new_next_bi;
+    }
+  else
+    {
+      if (~0 != reass->first_bi)
+	{
+	  new_next_vnb->ip.reass.next_range_bi = reass->first_bi;
+	}
+      reass->first_bi = new_next_bi;
+    }
+  reass->data_len += ip4_reass_buffer_get_data_len (new_next_b);
+  rm->buffers_n += ip4_reass_get_buffer_chain_length (vm, new_next_b);
+}
+
+static void
+ip4_reass_remove_range_from_chain (vlib_main_t * vm,
+				   vlib_node_runtime_t * node,
+				   ip4_reass_main_t * rm,
+				   u32 ** vec_drop_overlap,
+				   ip4_reass_t * reass, u32 prev_range_bi,
+				   u32 discard_bi)
+{
+  vlib_buffer_t *discard_b = vlib_get_buffer (vm, discard_bi);
+  vnet_buffer_opaque_t *discard_vnb = vnet_buffer (discard_b);
+  if (~0 != prev_range_bi)
+    {
+      vlib_buffer_t *prev_b = vlib_get_buffer (vm, prev_range_bi);
+      vnet_buffer_opaque_t *prev_vnb = vnet_buffer (prev_b);
+      ASSERT (prev_vnb->ip.reass.next_range_bi == discard_bi);
+      prev_vnb->ip.reass.next_range_bi = discard_vnb->ip.reass.next_range_bi;
+    }
+  else
+    {
+      reass->first_bi = discard_vnb->ip.reass.next_range_bi;
+    }
+  reass->data_len -= ip4_reass_buffer_get_data_len (discard_b);
+  while (1)
+    {
+      vec_add1 (*vec_drop_overlap, discard_bi);
+      if (PREDICT_FALSE (discard_b->flags & VLIB_BUFFER_IS_TRACED))
+	{
+	  ip4_reass_add_trace (vm, node, rm, reass, discard_bi, RANGE_DISCARD,
+			       0);
+	}
+      if (discard_b->flags & VLIB_BUFFER_NEXT_PRESENT)
+	{
+	  discard_b->flags &= ~VLIB_BUFFER_NEXT_PRESENT;
+	  discard_bi = discard_b->next_buffer;
+	  discard_b = vlib_get_buffer (vm, discard_bi);
+	}
+      else
+	{
+	  break;
+	}
+    }
+}
+
+void
+ip4_reass_update (vlib_main_t * vm, vlib_node_runtime_t * node,
+		  ip4_reass_main_t * rm, ip4_reass_t * reass, u32 * bi0,
+		  u32 * next0, vlib_error_t * error0,
+		  u32 ** vec_drop_overlap, u32 ** vec_drop_compress,
+		  u32 next_input, u32 next_drop)
+{
+  int consumed = 0;
+  vlib_buffer_t *fb = vlib_get_buffer (vm, *bi0);
+  ip4_header_t *fip = vlib_buffer_get_current (fb);
+  ASSERT (fb->current_length >= sizeof (*fip));
+  vnet_buffer_opaque_t *fvnb = vnet_buffer (fb);
+  u32 fragment_first = fvnb->ip.reass.fragment_first =
+    ip4_get_fragment_offset_bytes (fip);
+  u32 fragment_length =
+    clib_net_to_host_u16 (fip->length) - ip4_header_bytes (fip);
+  u32 fragment_last = fvnb->ip.reass.fragment_last =
+    fragment_first + fragment_length - 1;
+  int more_fragments = ip4_get_fragment_more (fip);
+  u32 candidate_range_bi = reass->first_bi;
+  u32 prev_range_bi = ~0;
+  fvnb->ip.reass.range_first = fragment_first;
+  fvnb->ip.reass.range_last = fragment_last;
+  fvnb->ip.reass.next_range_bi = ~0;
+  if (!more_fragments)
+    {
+      reass->last_packet_octet = fragment_last;
+    }
+  if (~0 == reass->first_bi)
+    {
+      // starting a new reassembly
+      ip4_reass_insert_range_in_chain (vm, rm, reass, prev_range_bi, *bi0);
+      if (PREDICT_FALSE (fb->flags & VLIB_BUFFER_IS_TRACED))
+	{
+	  ip4_reass_add_trace (vm, node, rm, reass, *bi0, RANGE_NEW, 0);
+	}
+      *bi0 = ~0;
+      fvnb->ip.reass.estimated_mtu = clib_net_to_host_u16 (fip->length);
+      return;
+    }
+  fvnb->ip.reass.estimated_mtu = clib_min (clib_net_to_host_u16 (fip->length),
+					   fvnb->ip.reass.estimated_mtu);
+  while (~0 != candidate_range_bi)
+    {
+      vlib_buffer_t *candidate_b = vlib_get_buffer (vm, candidate_range_bi);
+      vnet_buffer_opaque_t *candidate_vnb = vnet_buffer (candidate_b);
+      if (fragment_first > candidate_vnb->ip.reass.range_last)
+	{
+	  // this fragments starts after candidate range
+	  prev_range_bi = candidate_range_bi;
+	  candidate_range_bi = candidate_vnb->ip.reass.next_range_bi;
+	  if (candidate_vnb->ip.reass.range_last < fragment_last &&
+	      ~0 == candidate_range_bi)
+	    {
+	      // special case - this fragment falls beyond all known ranges
+	      ip4_reass_insert_range_in_chain (vm, rm, reass, prev_range_bi,
+					       *bi0);
+	      consumed = 1;
+	      break;
+	    }
+	  continue;
+	}
+      if (fragment_last < candidate_vnb->ip.reass.range_first)
+	{
+	  // this fragment ends before candidate range without any overlap
+	  ip4_reass_insert_range_in_chain (vm, rm, reass, prev_range_bi,
+					   *bi0);
+	  consumed = 1;
+	}
+      else
+	{
+	  if (fragment_first >= candidate_vnb->ip.reass.range_first &&
+	      fragment_last <= candidate_vnb->ip.reass.range_last)
+	    {
+	      // this fragment is a (sub)part of existing range, ignore it
+	      if (PREDICT_FALSE (fb->flags & VLIB_BUFFER_IS_TRACED))
+		{
+		  ip4_reass_add_trace (vm, node, rm, reass, *bi0,
+				       RANGE_OVERLAP, 0);
+		}
+	      break;
+	    }
+	  int discard_candidate = 0;
+	  if (fragment_first < candidate_vnb->ip.reass.range_first)
+	    {
+	      u32 overlap =
+		fragment_last - candidate_vnb->ip.reass.range_first + 1;
+	      if (overlap < ip4_reass_buffer_get_data_len (candidate_b))
+		{
+		  candidate_vnb->ip.reass.range_first += overlap;
+		  ASSERT (reass->data_len >= overlap);
+		  reass->data_len -= overlap;
+		  if (PREDICT_FALSE (fb->flags & VLIB_BUFFER_IS_TRACED))
+		    {
+		      ip4_reass_add_trace (vm, node, rm, reass,
+					   candidate_range_bi, RANGE_SHRINK,
+					   overlap);
+		    }
+		  ip4_reass_insert_range_in_chain (vm, rm, reass,
+						   prev_range_bi, *bi0);
+		  consumed = 1;
+		}
+	      else
+		{
+		  discard_candidate = 1;
+		}
+	    }
+	  else if (fragment_last > candidate_vnb->ip.reass.range_last)
+	    {
+	      u32 overlap =
+		candidate_vnb->ip.reass.range_last - fragment_first + 1;
+	      if (overlap < ip4_reass_buffer_get_data_len (candidate_b))
+		{
+		  fvnb->ip.reass.range_first += overlap;
+		  if (~0 != candidate_vnb->ip.reass.next_range_bi)
+		    {
+		      prev_range_bi = candidate_range_bi;
+		      candidate_range_bi =
+			candidate_vnb->ip.reass.next_range_bi;
+		      continue;
+		    }
+		  else
+		    {
+		      // special case - last range discarded
+		      ip4_reass_insert_range_in_chain (vm, rm, reass,
+						       candidate_range_bi,
+						       *bi0);
+		      consumed = 1;
+		    }
+		}
+	      else
+		{
+		  discard_candidate = 1;
+		}
+	    }
+	  else
+	    {
+	      discard_candidate = 1;
+	    }
+	  if (discard_candidate)
+	    {
+	      u32 next_range_bi = candidate_vnb->ip.reass.next_range_bi;
+	      // discard candidate range, probe next range
+	      ip4_reass_remove_range_from_chain (vm, node, rm,
+						 vec_drop_overlap, reass,
+						 prev_range_bi,
+						 candidate_range_bi);
+	      if (~0 != next_range_bi)
+		{
+		  candidate_range_bi = next_range_bi;
+		  continue;
+		}
+	      else
+		{
+		  // special case - last range discarded
+		  ip4_reass_insert_range_in_chain (vm, rm, reass,
+						   prev_range_bi, *bi0);
+		  consumed = 1;
+		}
+	    }
+	}
+      break;
+    }
+  if (consumed)
+    {
+      if (PREDICT_FALSE (fb->flags & VLIB_BUFFER_IS_TRACED))
+	{
+	  ip4_reass_add_trace (vm, node, rm, reass, *bi0, RANGE_NEW, 0);
+	}
+    }
+  if (~0 != reass->last_packet_octet &&
+      reass->data_len == reass->last_packet_octet + 1)
+    {
+      ip4_reass_finalize (vm, node, rm, reass, bi0, next0, error0, next_input,
+			  vec_drop_compress, vec_drop_overlap);
+    }
+  else
+    {
+      if (consumed)
+	{
+	  *bi0 = ~0;
+	}
+      else
+	{
+	  *next0 = next_drop;
+	  *error0 = IP4_ERROR_REASS_DUPLICATE_FRAGMENT;
+	}
+    }
+}
+
+always_inline uword
+ip4_reassembly (vlib_main_t * vm, vlib_node_runtime_t * node,
+		vlib_frame_t * frame)
+{
+  u32 *from = vlib_frame_vector_args (frame);
+  u32 n_left_from, n_left_to_next, *to_next, next_index;
+  ip4_reass_main_t *rm = &ip4_reass_main;
+
+  n_left_from = frame->n_vectors;
+  next_index = node->cached_next_index;
+  static u32 *vec_drop_timeout = NULL;	// indexes of buffers which timed out
+  static u32 *vec_drop_overlap = NULL;	// indexes of buffers which were discarded due to overlap
+  static u32 *vec_drop_compress = NULL;	// indexes of buffers dicarded due to buffer compression
+  while (n_left_from > 0 || vec_len (vec_drop_timeout) > 0 ||
+	 vec_len (vec_drop_overlap) > 0)
+    {
+      vlib_get_next_frame (vm, node, next_index, to_next, n_left_to_next);
+
+      while (vec_len (vec_drop_timeout) > 0 && n_left_to_next > 0)
+	{
+	  u32 bi = vec_pop (vec_drop_timeout);
+	  vlib_buffer_t *b = vlib_get_buffer (vm, bi);
+	  b->error = node->errors[IP4_ERROR_REASS_TIMEOUT];
+	  to_next[0] = bi;
+	  to_next += 1;
+	  n_left_to_next -= 1;
+	  vlib_validate_buffer_enqueue_x1 (vm, node, next_index, to_next,
+					   n_left_to_next, bi,
+					   IP4_REASSEMBLY_NEXT_DROP);
+	  IP4_REASS_DEBUG_BUFFER (bi, enqueue_drop_timeout);
+	  ASSERT (rm->buffers_n > 0);
+	  --rm->buffers_n;
+	}
+
+      while (vec_len (vec_drop_overlap) > 0 && n_left_to_next > 0)
+	{
+	  u32 bi = vec_pop (vec_drop_overlap);
+	  vlib_buffer_t *b = vlib_get_buffer (vm, bi);
+	  b->error = node->errors[IP4_ERROR_REASS_DUPLICATE_FRAGMENT];
+	  to_next[0] = bi;
+	  to_next += 1;
+	  n_left_to_next -= 1;
+	  vlib_validate_buffer_enqueue_x1 (vm, node, next_index, to_next,
+					   n_left_to_next, bi,
+					   IP4_REASSEMBLY_NEXT_DROP);
+	  IP4_REASS_DEBUG_BUFFER (bi, enqueue_drop_duplicate_fragment);
+	  ASSERT (rm->buffers_n > 0);
+	  --rm->buffers_n;
+	}
+
+      while (vec_len (vec_drop_compress) > 0 && n_left_to_next > 0)
+	{
+	  u32 bi = vec_pop (vec_drop_compress);
+	  vlib_buffer_t *b = vlib_get_buffer (vm, bi);
+	  b->error = node->errors[IP4_ERROR_NONE];
+	  to_next[0] = bi;
+	  to_next += 1;
+	  n_left_to_next -= 1;
+	  vlib_validate_buffer_enqueue_x1 (vm, node, next_index, to_next,
+					   n_left_to_next, bi,
+					   IP4_REASSEMBLY_NEXT_DROP);
+	  IP4_REASS_DEBUG_BUFFER (bi, enqueue_drop_compress);
+	  ASSERT (rm->buffers_n > 0);
+	  --rm->buffers_n;
+	}
+
+      while (n_left_from > 0 && n_left_to_next > 0)
+	{
+	  u32 bi0;
+	  vlib_buffer_t *b0;
+	  u32 next0;		//, error0;
+
+	  bi0 = from[0];
+	  b0 = vlib_get_buffer (vm, bi0);
+
+	  ip4_header_t *ip0 = vlib_buffer_get_current (b0);
+	  ip4_reass_key_t k;
+	  k.src.as_u32 = ip0->src_address.as_u32;
+	  k.dst.as_u32 = ip0->dst_address.as_u32;
+	  k.xx_id = vnet_buffer (b0)->sw_if_index[VLIB_RX];
+	  k.frag_id = ip0->fragment_id;
+	  k.proto = ip0->protocol;
+	  k.unused = 0;
+	  ip4_reass_t *reass =
+	    ip4_reass_find_or_create (vm, rm, &k, &vec_drop_timeout);
+
+	  u32 error0 = IP4_ERROR_NONE;
+	  if (reass)
+	    {
+	      ip4_reass_update (vm, node, rm, reass, &bi0, &next0, &error0,
+				&vec_drop_overlap, &vec_drop_compress,
+				IP4_REASSEMBLY_NEXT_INPUT,
+				IP4_REASSEMBLY_NEXT_DROP);
+	    }
+	  else
+	    {
+	      next0 = IP4_REASSEMBLY_NEXT_DROP;
+	      error0 = IP4_ERROR_REASS_LIMIT_REACHED;
+	    }
+
+	  b0->error = node->errors[error0];
+
+	  if (bi0 != ~0)
+	    {
+	      to_next[0] = bi0;
+	      to_next += 1;
+	      n_left_to_next -= 1;
+	      vlib_validate_buffer_enqueue_x1 (vm, node, next_index, to_next,
+					       n_left_to_next, bi0, next0);
+	      IP4_REASS_DEBUG_BUFFER (bi0, enqueue_next);
+	    }
+
+	  from += 1;
+	  n_left_from -= 1;
+	}
+
+      vlib_put_next_frame (vm, node, next_index, n_left_to_next);
+    }
+
+  return frame->n_vectors;
+}
+
+static char *ip4_reassembly_error_strings[] = {
+#define _(sym, string) string,
+  foreach_ip4_error
+#undef _
+};
+
+/* *INDENT-OFF* */
+VLIB_REGISTER_NODE (ip4_reass_node, static) = {
+    .function = ip4_reassembly,
+    .name = "ip4-reassembly",
+    .vector_size = sizeof (u32),
+    .format_trace = format_ip4_reass_trace,
+    .n_errors = ARRAY_LEN (ip4_reassembly_error_strings),
+    .error_strings = ip4_reassembly_error_strings,
+    .n_next_nodes = IP4_REASSEMBLY_N_NEXT,
+    .next_nodes =
+        {
+                [IP4_REASSEMBLY_NEXT_INPUT] = "ip4-input",
+                [IP4_REASSEMBLY_NEXT_DROP] = "ip4-drop",
+        },
+};
+/* *INDENT-ON* */
+
+VLIB_NODE_FUNCTION_MULTIARCH (ip4_reass_node, ip4_reassembly)
+     static u32 ip4_reass_get_nbuckets ()
+{
+  ip4_reass_main_t *rm = &ip4_reass_main;
+  u32 nbuckets;
+  u8 i;
+
+  nbuckets = (u32) (rm->max_reass_n / IP4_REASS_HT_LOAD_FACTOR);
+
+  for (i = 0; i < 31; i++)
+    if ((1 << i) >= nbuckets)
+      break;
+  nbuckets = 1 << i;
+
+  return nbuckets;
+}
+
+typedef enum
+{
+  IP4_EVENT_CONFIG_CHANGED = 1,
+} ip4_reass_event_t;
+
+typedef struct
+{
+  int failure;
+  clib_bihash_24_8_t *new_hash;
+} ip4_rehash_cb_ctx;
+
+void
+ip4_rehash_cb (clib_bihash_kv_24_8_t * kv, void *_ctx)
+{
+  ip4_rehash_cb_ctx *ctx = _ctx;
+  if (clib_bihash_add_del_24_8 (ctx->new_hash, kv, 1))
+    {
+      ctx->failure = 1;
+    }
+}
+
+vnet_api_error_t
+ip4_reass_set (u32 timeout_ms, u32 max_reassemblies,
+	       u32 expire_walk_interval_ms)
+{
+  u32 old_nbuckets = ip4_reass_get_nbuckets ();
+  ip4_reass_main.timeout_ms = timeout_ms;
+  ip4_reass_main.timeout = (f64) timeout_ms / (f64) MSEC_PER_SEC;
+  ip4_reass_main.max_reass_n = max_reassemblies;
+  ip4_reass_main.expire_walk_interval_ms = expire_walk_interval_ms;
+  vlib_process_signal_event (ip4_reass_main.vlib_main,
+			     ip4_reass_main.ip4_reass_expire_node_idx,
+			     IP4_EVENT_CONFIG_CHANGED, 0);
+  u32 new_nbuckets = ip4_reass_get_nbuckets ();
+  if (ip4_reass_main.max_reass_n > 0 && new_nbuckets > 1 &&
+      new_nbuckets != old_nbuckets)
+    {
+      clib_bihash_24_8_t new_hash;
+      memset (&new_hash, 0, sizeof (new_hash));
+      ip4_rehash_cb_ctx ctx;
+      ctx.failure = 0;
+      ctx.new_hash = &new_hash;
+      clib_bihash_init_24_8 (&new_hash, "ip4-reass", new_nbuckets,
+			     new_nbuckets * 1024);
+      clib_bihash_foreach_key_value_pair_24_8 (&ip4_reass_main.hash,
+					       ip4_rehash_cb, &ctx);
+      if (ctx.failure)
+	{
+	  clib_bihash_free_24_8 (&new_hash);
+	  return -1;
+	}
+      else
+	{
+	  clib_bihash_free_24_8 (&ip4_reass_main.hash);
+	  clib_memcpy (&ip4_reass_main.hash, &new_hash,
+		       sizeof (ip4_reass_main.hash));
+	}
+    }
+  return 0;
+}
+
+vnet_api_error_t
+ip4_reass_get (u32 * timeout_ms, u32 * max_reassemblies,
+	       u32 * expire_walk_interval_ms)
+{
+  *timeout_ms = ip4_reass_main.timeout_ms;
+  *max_reassemblies = ip4_reass_main.max_reass_n;
+  *expire_walk_interval_ms = ip4_reass_main.expire_walk_interval_ms;
+  return 0;
+}
+
+clib_error_t *
+ip4_reass_init_function (vlib_main_t * vm)
+{
+  ip4_reass_main_t *rm = &ip4_reass_main;
+  clib_error_t *error = 0;
+  u32 nbuckets;
+
+  rm->vlib_main = vm;
+  rm->vnet_main = vnet_get_main ();
+
+  rm->reass_n = 0;
+  pool_alloc (rm->pool, rm->max_reass_n);
+  ip4_reass_set (IP4_REASS_TIMEOUT_DEFAULT_MS,
+		 IP4_REASS_MAX_REASSEMBLIES_DEAFULT,
+		 IP4_REASS_EXPIRE_WALK_INTERVAL_DEFAULT_MS);
+
+  nbuckets = ip4_reass_get_nbuckets ();
+  clib_bihash_init_24_8 (&rm->hash, "ip4-reass", nbuckets, nbuckets * 1024);
+
+  vlib_node_t *node = vlib_get_node_by_name (vm, (u8 *) "ip4-drop");
+  ASSERT (node);
+  rm->ip4_drop_idx = node->index;
+  node = vlib_get_node_by_name (vm, (u8 *) "ip4-reassembly-expire-walk");
+  ASSERT (node);
+  rm->ip4_reass_expire_node_idx = node->index;
+  return error;
+}
+
+VLIB_INIT_FUNCTION (ip4_reass_init_function);
+
+static uword
+ip4_reass_walk_expired (vlib_main_t * vm,
+			vlib_node_runtime_t * node, vlib_frame_t * f)
+{
+  ip4_reass_main_t *rm = &ip4_reass_main;
+  uword event_type, *event_data = 0;
+
+  while (true)
+    {
+      vlib_process_wait_for_event_or_clock (vm,
+					    (f64) rm->expire_walk_interval_ms
+					    / (f64) MSEC_PER_SEC);
+      event_type = vlib_process_get_events (vm, &event_data);
+
+      switch (event_type)
+	{
+	case ~0:		/* no events => timeout */
+	  /* nothing to do here */
+	  break;
+	case IP4_EVENT_CONFIG_CHANGED:
+	  break;
+	default:
+	  clib_warning ("BUG: event type 0x%wx", event_type);
+	  break;
+	}
+      f64 now = vlib_time_now (vm);
+
+      ip4_reass_t *reass;
+      u32 *vec_drop_timeout = NULL;
+      int *pool_indexes_to_free = NULL;
+
+      int index;
+      /* *INDENT-OFF* */
+      pool_foreach_index (index, rm->pool, ({
+                            reass = pool_elt_at_index (rm->pool, index);
+                            if (now > reass->last_heard + rm->timeout)
+                              {
+                                vec_add1 (pool_indexes_to_free, index);
+                              }
+                          }));
+      /* *INDENT-ON* */
+      int *i;
+      /* *INDENT-OFF* */
+      vec_foreach (i, pool_indexes_to_free)
+      {
+        ip4_reass_t *reass = pool_elt_at_index (rm->pool, i[0]);
+        ip4_reass_on_timeout (vm, rm, reass, &vec_drop_timeout);
+        ip4_reass_free (rm, reass);
+      }
+      /* *INDENT-ON* */
+
+      while (vec_len (vec_drop_timeout) > 0)
+	{
+	  vlib_frame_t *f = vlib_get_frame_to_node (vm, rm->ip4_drop_idx);
+	  u32 *to_next = vlib_frame_vector_args (f);
+	  u32 n_left_to_next = VLIB_FRAME_SIZE - f->n_vectors;
+	  u32 n_trace = 0;
+	  while (vec_len (vec_drop_timeout) > 0 && n_left_to_next > 0)
+	    {
+	      u32 bi = vec_pop (vec_drop_timeout);
+	      vlib_buffer_t *b = vlib_get_buffer (vm, bi);
+	      if (PREDICT_FALSE (b->flags & VLIB_BUFFER_IS_TRACED))
+		{
+		  if (pool_is_free_index (vm->trace_main.trace_buffer_pool,
+					  b->trace_index))
+		    {
+		      /* the trace is gone, don't trace this buffer anymore */
+		      b->flags &= ~VLIB_BUFFER_IS_TRACED;
+		    }
+		  else
+		    {
+		      ++n_trace;
+		    }
+		}
+	      b->error = node->errors[IP4_ERROR_REASS_TIMEOUT];
+	      to_next[0] = bi;
+	      ++f->n_vectors;
+	      to_next += 1;
+	      n_left_to_next -= 1;
+	      IP4_REASS_DEBUG_BUFFER (bi, enqueue_drop_timeout_walk);
+	      ASSERT (rm->buffers_n > 0);
+	      --rm->buffers_n;
+	    }
+	  if (PREDICT_FALSE (n_trace > 0))
+	    {
+	      f->flags |= VLIB_FRAME_TRACE;
+	    }
+	  vlib_put_frame_to_node (vm, rm->ip4_drop_idx, f);
+	}
+
+      vec_free (pool_indexes_to_free);
+      vec_free (vec_drop_timeout);
+      if (event_data)
+	{
+	  _vec_len (event_data) = 0;
+	}
+    }
+
+  return 0;
+}
+
+static vlib_node_registration_t ip4_reass_expire_node;
+
+/* *INDENT-OFF* */
+VLIB_REGISTER_NODE (ip4_reass_expire_node, static) = {
+    .function = ip4_reass_walk_expired,
+    .type = VLIB_NODE_TYPE_PROCESS,
+    .name = "ip4-reassembly-expire-walk",
+    .format_trace = format_ip4_reass_trace,
+    .n_errors = ARRAY_LEN (ip4_reassembly_error_strings),
+    .error_strings = ip4_reassembly_error_strings,
+
+};
+/* *INDENT-ON* */
+
+static u8 *
+format_ip4_reass_key (u8 * s, va_list * args)
+{
+  ip4_reass_key_t *key = va_arg (*args, ip4_reass_key_t *);
+  s = format (s, "xx_id: %u, src: %U, dst: %U, frag_id: %u, proto: %u",
+	      key->xx_id, format_ip4_address, &key->src, format_ip4_address,
+	      &key->dst, clib_net_to_host_u16 (key->frag_id), key->proto);
+  return s;
+}
+
+static u8 *
+format_ip4_reass (u8 * s, va_list * args)
+{
+  vlib_main_t *vm = va_arg (*args, vlib_main_t *);
+  ip4_reass_t *reass = va_arg (*args, ip4_reass_t *);
+
+  s = format (s, "ID: %u, key: %U\n  first_bi: %u, data_len: %u, "
+	      "last_packet_octet: %u, trace_op_counter: %u\n",
+	      reass->id, format_ip4_reass_key, &reass->key, reass->first_bi,
+	      reass->data_len, reass->last_packet_octet,
+	      reass->trace_op_counter);
+  u32 bi = reass->first_bi;
+  u32 counter = 0;
+  while (~0 != bi)
+    {
+      vlib_buffer_t *b = vlib_get_buffer (vm, bi);
+      vnet_buffer_opaque_t *vnb = vnet_buffer (b);
+      s = format (s, "  #%03u: range: [%u, %u], bi: %u, off: %d, len: %u, "
+		  "fragment[%u, %u]\n",
+		  counter, vnb->ip.reass.range_first,
+		  vnb->ip.reass.range_last, bi,
+		  ip4_reass_buffer_get_data_offset_no_check (b),
+		  ip4_reass_buffer_get_data_len_no_check (b),
+		  vnb->ip.reass.fragment_first, vnb->ip.reass.fragment_last);
+      if (b->flags & VLIB_BUFFER_NEXT_PRESENT)
+	{
+	  bi = b->next_buffer;
+	}
+      else
+	{
+	  bi = ~0;
+	}
+    }
+  return s;
+}
+
+static clib_error_t *
+show_ip4_reass (vlib_main_t * vm, unformat_input_t * input,
+		CLIB_UNUSED (vlib_cli_command_t * lmd))
+{
+  ip4_reass_main_t *rm = &ip4_reass_main;
+
+  vlib_cli_output (vm, "---------------------");
+  vlib_cli_output (vm, "IP4 reassembly status");
+  vlib_cli_output (vm, "---------------------");
+  if (unformat (input, "details"))
+    {
+      ip4_reass_t *reass;
+      /* *INDENT-OFF* */
+      pool_foreach (reass, rm->pool, {
+        vlib_cli_output (vm, "%U", format_ip4_reass, vm, reass);
+      });
+      /* *INDENT-ON* */
+    }
+  vlib_cli_output (vm, "---------------------");
+  vlib_cli_output (vm, "Current IP4 reassemblies count: %lu\n", rm->reass_n);
+  vlib_cli_output (vm,
+		   "Maximum configured concurrent IP4 reassemblies: %lu\n",
+		   (long unsigned) rm->max_reass_n);
+  vlib_cli_output (vm, "Buffers in use: %lu\n",
+		   (long unsigned) rm->buffers_n);
+  return 0;
+}
+
+/* *INDENT-OFF* */
+VLIB_CLI_COMMAND (show_ip4_reassembly_cmd, static) = {
+    .path = "show ip4-reassembly",
+    .short_help = "show ip4-reassembly [details]",
+    .function = show_ip4_reass,
+};
+/* *INDENT-ON* */
+
+/*
+ * fd.io coding-style-patch-verification: ON
+ *
+ * Local Variables:
+ * eval: (c-set-style "gnu")
+ * End:
+ */
diff --git a/src/vnet/ip/ip4_reassembly.h b/src/vnet/ip/ip4_reassembly.h
new file mode 100644
index 00000000000..a48c07ef061
--- /dev/null
+++ b/src/vnet/ip/ip4_reassembly.h
@@ -0,0 +1,49 @@
+/*
+ * Copyright (c) 2017 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 IPv4 Reassembly.
+ *
+ * This file contains the source code for IPv4 reassembly.
+ */
+
+#ifndef __included_ip4_reassembly_h__
+#define __included_ip4_reassembly_h__
+
+#include <vnet/api_errno.h>
+#include <vnet/vnet.h>
+
+/**
+ * @brief set ip4 reassembly configuration
+ */
+vnet_api_error_t ip4_reass_set (u32 timeout_ms, u32 max_reassemblies,
+				u32 expire_walk_interval_ms);
+
+/**
+ * @brief get ip4 reassembly configuration
+ */
+vnet_api_error_t ip4_reass_get (u32 * timeout_ms, u32 * max_reassemblies,
+				u32 * expire_walk_interval_ms);
+
+#endif /* __included_ip4_reassembly_h */
+
+/*
+ * fd.io coding-style-patch-verification: ON
+ *
+ * Local Variables:
+ * eval: (c-set-style "gnu")
+ * End:
+ */
diff --git a/src/vnet/ip/ip6_error.h b/src/vnet/ip/ip6_error.h
index 7bb4b77b5a3..19f20625a10 100644
--- a/src/vnet/ip/ip6_error.h
+++ b/src/vnet/ip/ip6_error.h
@@ -71,7 +71,14 @@
                                                                         \
  /* Erros singalled by ip6-inacl */                                     \
   _ (INACL_TABLE_MISS, "input ACL table-miss drops")                    \
-  _ (INACL_SESSION_DENY, "input ACL session deny drops")
+  _ (INACL_SESSION_DENY, "input ACL session deny drops")                \
+                                                                        \
+  /* Errors signalled by ip6-reassembly */                              \
+  _ (REASS_MISSING_UPPER, "missing-upper layer drops")                  \
+  _ (REASS_DUPLICATE_FRAGMENT, "duplicate fragments")                   \
+  _ (REASS_OVERLAPPING_FRAGMENT, "overlapping fragments")               \
+  _ (REASS_LIMIT_REACHED, "drops due to concurrent reassemblies limit") \
+  _ (REASS_TIMEOUT, "fragments dropped due to reassembly timeout")
 
 typedef enum
 {
diff --git a/src/vnet/ip/ip6_forward.c b/src/vnet/ip/ip6_forward.c
index c1c9ec0b586..5b06ca7609d 100644
--- a/src/vnet/ip/ip6_forward.c
+++ b/src/vnet/ip/ip6_forward.c
@@ -1578,6 +1578,7 @@ ip6_local_inline (vlib_main_t * vm, vlib_node_runtime_t * node,
 	  next0 = lm->local_next_by_ip_protocol[ip0->protocol];
 	  next0 =
 	    error0 != IP6_ERROR_UNKNOWN_PROTOCOL ? IP_LOCAL_NEXT_DROP : next0;
+
 	  p0->error = error_node->errors[error0];
 
 	  if (head_of_feature_arc)
@@ -1619,6 +1620,7 @@ VLIB_REGISTER_NODE (ip6_local_node, static) =
     [IP_LOCAL_NEXT_PUNT] = "ip6-punt",
     [IP_LOCAL_NEXT_UDP_LOOKUP] = "ip6-udp-lookup",
     [IP_LOCAL_NEXT_ICMP] = "ip6-icmp-input",
+    [IP_LOCAL_NEXT_REASSEMBLY] = "ip6-reassembly",
   },
 };
 /* *INDENT-ON* */
diff --git a/src/vnet/ip/ip6_input.h b/src/vnet/ip/ip6_input.h
index 4c0d78459b1..fe993caa889 100644
--- a/src/vnet/ip/ip6_input.h
+++ b/src/vnet/ip/ip6_input.h
@@ -41,6 +41,7 @@
 #define included_ip6_input_h
 
 #include <vnet/ip/ip.h>
+#include <vnet/ip/icmp6.h>
 
 extern char *ip6_error_strings[];
 
diff --git a/src/vnet/ip/ip6_packet.h b/src/vnet/ip/ip6_packet.h
index c0c745e27fe..6b724370e76 100644
--- a/src/vnet/ip/ip6_packet.h
+++ b/src/vnet/ip/ip6_packet.h
@@ -519,6 +519,9 @@ typedef CLIB_PACKED (struct {
 #define ip6_frag_hdr_offset(hdr) \
   (clib_net_to_host_u16((hdr)->fragment_offset_and_more) >> 3)
 
+#define ip6_frag_hdr_offset_bytes(hdr) \
+  (8 * ip6_frag_hdr_offset(hdr))
+
 #define ip6_frag_hdr_more(hdr) \
   (clib_net_to_host_u16((hdr)->fragment_offset_and_more) & 0x1)
 
diff --git a/src/vnet/ip/ip6_reassembly.c b/src/vnet/ip/ip6_reassembly.c
new file mode 100644
index 00000000000..d448032b913
--- /dev/null
+++ b/src/vnet/ip/ip6_reassembly.c
@@ -0,0 +1,1347 @@
+/*
+ * Copyright (c) 2017 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 IPv6 Reassembly.
+ *
+ * This file contains the source code for IPv6 reassembly.
+ */
+
+#include <vppinfra/vec.h>
+#include <vnet/vnet.h>
+#include <vnet/ip/ip.h>
+#include <vppinfra/bihash_48_8.h>
+#include <vnet/ip/ip6_reassembly.h>
+
+#define MSEC_PER_SEC 1000
+#define IP6_REASS_TIMEOUT_DEFAULT_MS 100
+#define IP6_REASS_EXPIRE_WALK_INTERVAL_DEFAULT_MS 10000	// 10 seconds default
+#define IP6_REASS_MAX_REASSEMBLIES_DEAFULT 1024
+#define IP6_REASS_HT_LOAD_FACTOR (0.75)
+
+static vlib_node_registration_t ip6_reass_node;
+
+typedef struct
+{
+  union
+  {
+    struct
+    {
+      ip6_address_t src;
+      ip6_address_t dst;
+      // align by making this 4 octets even though its a 2 octets field
+      u32 xx_id;
+      // align by making this 4 octets even though its a 2 octets field
+      u32 frag_id;
+      // align by making this 4 octets even though its a 1 octet field
+      u32 proto;
+      u32 unused;
+    };
+    u64 as_u64[6];
+  };
+} ip6_reass_key_t;
+
+always_inline u32
+ip6_reass_buffer_get_data_offset_no_check (vlib_buffer_t * b)
+{
+  vnet_buffer_opaque_t *vnb = vnet_buffer (b);
+  return vnb->ip.reass.range_first - vnb->ip.reass.fragment_first;
+}
+
+always_inline u32
+ip6_reass_buffer_get_data_offset (vlib_buffer_t * b)
+{
+  vnet_buffer_opaque_t *vnb = vnet_buffer (b);
+  ASSERT (vnb->ip.reass.range_first >= vnb->ip.reass.fragment_first);
+  return ip6_reass_buffer_get_data_offset_no_check (b);
+}
+
+always_inline u16
+ip6_reass_buffer_get_data_len_no_check (vlib_buffer_t * b)
+{
+  vnet_buffer_opaque_t *vnb = vnet_buffer (b);
+  return clib_min (vnb->ip.reass.range_last, vnb->ip.reass.fragment_last) -
+    (vnb->ip.reass.fragment_first + ip6_reass_buffer_get_data_offset (b)) + 1;
+}
+
+always_inline u16
+ip6_reass_buffer_get_data_len (vlib_buffer_t * b)
+{
+  vnet_buffer_opaque_t *vnb = vnet_buffer (b);
+  ASSERT (vnb->ip.reass.range_last > vnb->ip.reass.fragment_first);
+  return ip6_reass_buffer_get_data_len_no_check (b);
+}
+
+typedef struct
+{
+  // hash table key
+  ip6_reass_key_t key;
+  // time when first packet was received
+  f64 first_heard;
+  // time when last packet was received
+  f64 last_heard;
+  // internal id of this reassembly
+  u32 id;
+  // buffer index of first buffer in this reassembly context
+  u32 first_bi;
+  // last octet of packet, ~0 until fragment without more_fragments arrives
+  u32 last_packet_octet;
+  // length of data collected so far
+  u32 data_len;
+  // trace operation counter
+  u32 trace_op_counter;
+} ip6_reass_t;
+
+typedef struct
+{
+  // IPv6 config
+  u32 timeout_ms;
+  f64 timeout;
+  u32 expire_walk_interval_ms;
+  u32 max_reass_n;
+
+  // IPv6 runtime
+  ip6_reass_t *pool;
+  clib_bihash_48_8_t hash;
+  u32 reass_n;
+  u32 id_counter;
+  u32 buffers_n;
+
+  // convenience
+  vlib_main_t *vlib_main;
+  vnet_main_t *vnet_main;
+
+  // node index of ip6-drop node
+  u32 ip6_drop_idx;
+  u32 ip6_icmp_error_idx;
+  u32 ip6_reass_expire_node_idx;
+
+} ip6_reass_main_t;
+
+ip6_reass_main_t ip6_reass_main;
+
+typedef enum
+{
+  IP6_REASSEMBLY_NEXT_INPUT,
+  IP6_REASSEMBLY_NEXT_DROP,
+  IP6_REASSEMBLY_NEXT_ICMP_ERROR,
+  IP6_REASSEMBLY_N_NEXT,
+} ip6_reass_next_t;
+
+typedef enum
+{
+  RANGE_NEW,
+  RANGE_OVERLAP,
+  ICMP_ERROR_RT_EXCEEDED,
+  ICMP_ERROR_FL_TOO_BIG,
+  ICMP_ERROR_FL_NOT_MULT_8,
+  FINALIZE,
+} ip6_reass_trace_operation_e;
+
+typedef struct
+{
+  u16 range_first;
+  u16 range_last;
+  u32 range_bi;
+  i32 data_offset;
+  u32 data_len;
+  u32 first_bi;
+} ip6_reass_range_trace_t;
+
+typedef struct
+{
+  ip6_reass_trace_operation_e action;
+  u32 pool_index;
+  u32 reass_id;
+  ip6_reass_range_trace_t trace_range;
+  u32 size_diff;
+  u32 op_id;
+  u32 fragment_first;
+  u32 fragment_last;
+  u32 total_data_len;
+} ip6_reass_trace_t;
+
+static void
+ip6_reass_trace_details (vlib_main_t * vm, u32 bi,
+			 ip6_reass_range_trace_t * trace)
+{
+  vlib_buffer_t *b = vlib_get_buffer (vm, bi);
+  vnet_buffer_opaque_t *vnb = vnet_buffer (b);
+  trace->range_first = vnb->ip.reass.range_first;
+  trace->range_last = vnb->ip.reass.range_last;
+  trace->data_offset = ip6_reass_buffer_get_data_offset_no_check (b);
+  trace->data_len = ip6_reass_buffer_get_data_len_no_check (b);
+  trace->range_bi = bi;
+}
+
+u8 *
+format_ip6_reass_range_trace (u8 * s, va_list * args)
+{
+  ip6_reass_range_trace_t *trace = va_arg (*args, ip6_reass_range_trace_t *);
+  s = format (s, "range: [%u, %u], off %d, len %u, bi %u", trace->range_first,
+	      trace->range_last, trace->data_offset, trace->data_len,
+	      trace->range_bi);
+  return s;
+}
+
+u8 *
+format_ip6_reass_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 *);
+  ip6_reass_trace_t *t = va_arg (*args, ip6_reass_trace_t *);
+  s = format (s, "reass id: %u, op id: %u ", t->reass_id, t->op_id);
+  u32 indent = format_get_indent (s);
+  s = format (s, "first bi: %u, data len: %u, ip/fragment[%u, %u]",
+	      t->trace_range.first_bi, t->total_data_len, t->fragment_first,
+	      t->fragment_last);
+  switch (t->action)
+    {
+    case RANGE_NEW:
+      s = format (s, "\n%Unew %U", format_white_space, indent,
+		  format_ip6_reass_range_trace, &t->trace_range);
+      break;
+    case RANGE_OVERLAP:
+      s = format (s, "\n%Uoverlap %U", format_white_space, indent,
+		  format_ip6_reass_range_trace, &t->trace_range);
+      break;
+    case ICMP_ERROR_FL_TOO_BIG:
+      s = format (s, "\n%Uicmp-error - frag_len > 65535 %U",
+		  format_white_space, indent, format_ip6_reass_range_trace,
+		  &t->trace_range);
+      break;
+    case ICMP_ERROR_FL_NOT_MULT_8:
+      s = format (s, "\n%Uicmp-error - frag_len mod 8 != 0 %U",
+		  format_white_space, indent, format_ip6_reass_range_trace,
+		  &t->trace_range);
+      break;
+    case ICMP_ERROR_RT_EXCEEDED:
+      s = format (s, "\n%Uicmp-error - reassembly time exceeded",
+		  format_white_space, indent);
+      break;
+    case FINALIZE:
+      s = format (s, "\n%Ufinalize reassembly", format_white_space, indent);
+      break;
+    }
+  return s;
+}
+
+static void
+ip6_reass_add_trace (vlib_main_t * vm, vlib_node_runtime_t * node,
+		     ip6_reass_main_t * rm, ip6_reass_t * reass,
+		     u32 bi, ip6_reass_trace_operation_e action,
+		     u32 size_diff)
+{
+  vlib_buffer_t *b = vlib_get_buffer (vm, bi);
+  vnet_buffer_opaque_t *vnb = vnet_buffer (b);
+  ip6_reass_trace_t *t = vlib_add_trace (vm, node, b, sizeof (t[0]));
+  t->pool_index = reass - rm->pool;
+  t->reass_id = reass->id;
+  t->action = action;
+  ip6_reass_trace_details (vm, bi, &t->trace_range);
+  t->size_diff = size_diff;
+  t->op_id = reass->trace_op_counter;
+  ++reass->trace_op_counter;
+  t->fragment_first = vnb->ip.reass.fragment_first;
+  t->fragment_last = vnb->ip.reass.fragment_last;
+  t->trace_range.first_bi = reass->first_bi;
+  t->total_data_len = reass->data_len;
+#if 0
+  static u8 *s = NULL;
+  s = format (s, "%U", format_ip6_reass_trace, NULL, NULL, t);
+  printf ("%.*s\n", vec_len (s), s);
+  fflush (stdout);
+  vec_reset_length (s);
+#endif
+}
+
+static void
+ip6_reass_free (ip6_reass_main_t * rm, ip6_reass_t * reass)
+{
+  clib_bihash_kv_48_8_t kv;
+  kv.key[0] = reass->key.as_u64[0];
+  kv.key[1] = reass->key.as_u64[1];
+  kv.key[2] = reass->key.as_u64[2];
+  kv.key[3] = reass->key.as_u64[3];
+  kv.key[4] = reass->key.as_u64[4];
+  kv.key[5] = reass->key.as_u64[5];
+  clib_bihash_add_del_48_8 (&rm->hash, &kv, 0);
+  pool_put (rm->pool, reass);
+  --rm->reass_n;
+}
+
+static void
+ip6_reass_drop_all (vlib_main_t * vm, ip6_reass_main_t * rm,
+		    ip6_reass_t * reass, u32 ** vec_drop_bi)
+{
+  u32 range_bi = reass->first_bi;
+  vlib_buffer_t *range_b;
+  vnet_buffer_opaque_t *range_vnb;
+  while (~0 != range_bi)
+    {
+      range_b = vlib_get_buffer (vm, range_bi);
+      range_vnb = vnet_buffer (range_b);
+      u32 bi = range_bi;
+      while (~0 != bi)
+	{
+	  vec_add1 (*vec_drop_bi, bi);
+	  vlib_buffer_t *b = vlib_get_buffer (vm, bi);
+	  if (b->flags & VLIB_BUFFER_NEXT_PRESENT)
+	    {
+	      bi = b->next_buffer;
+	      b->flags &= ~VLIB_BUFFER_NEXT_PRESENT;
+	    }
+	  else
+	    {
+	      bi = ~0;
+	    }
+	}
+      range_bi = range_vnb->ip.reass.next_range_bi;
+    }
+}
+
+static void
+ip6_reass_on_timeout (vlib_main_t * vm, vlib_node_runtime_t * node,
+		      ip6_reass_main_t * rm, ip6_reass_t * reass,
+		      u32 * icmp_bi, u32 ** vec_timeout)
+{
+  if (~0 == reass->first_bi)
+    {
+      return;
+    }
+  vlib_buffer_t *b = vlib_get_buffer (vm, reass->first_bi);
+  if (0 == vnet_buffer (b)->ip.reass.fragment_first)
+    {
+      *icmp_bi = reass->first_bi;
+      if (PREDICT_FALSE (b->flags & VLIB_BUFFER_IS_TRACED))
+	{
+	  ip6_reass_add_trace (vm, node, rm, reass, reass->first_bi,
+			       ICMP_ERROR_RT_EXCEEDED, 0);
+	}
+      // fragment with offset zero received - send icmp message back
+      if (b->flags & VLIB_BUFFER_NEXT_PRESENT)
+	{
+	  // separate first buffer from chain and steer it towards icmp node
+	  b->flags &= ~VLIB_BUFFER_NEXT_PRESENT;
+	  reass->first_bi = b->next_buffer;
+	}
+      else
+	{
+	  reass->first_bi = vnet_buffer (b)->ip.reass.next_range_bi;
+	}
+      icmp6_error_set_vnet_buffer (b, ICMP6_time_exceeded,
+				   ICMP6_time_exceeded_fragment_reassembly_time_exceeded,
+				   0);
+    }
+  ip6_reass_drop_all (vm, rm, reass, vec_timeout);
+}
+
+static ip6_reass_t *
+ip6_reass_find_or_create (vlib_main_t * vm,
+			  vlib_node_runtime_t * node,
+			  ip6_reass_main_t * rm,
+			  ip6_reass_key_t * k, u32 * icmp_bi,
+			  u32 ** vec_timeout)
+{
+  ip6_reass_t *reass = NULL;
+  f64 now = vlib_time_now (rm->vlib_main);
+  clib_bihash_kv_48_8_t kv, value;
+  kv.key[0] = k->as_u64[0];
+  kv.key[1] = k->as_u64[1];
+  kv.key[2] = k->as_u64[2];
+  kv.key[3] = k->as_u64[3];
+  kv.key[4] = k->as_u64[4];
+  kv.key[5] = k->as_u64[5];
+
+  if (!clib_bihash_search_48_8 (&rm->hash, &kv, &value))
+    {
+      reass = pool_elt_at_index (rm->pool, value.value);
+      if (now > reass->last_heard + rm->timeout)
+	{
+	  ip6_reass_on_timeout (vm, node, rm, reass, icmp_bi, vec_timeout);
+	  ip6_reass_free (rm, reass);
+	  reass = NULL;
+	}
+    }
+
+  if (reass)
+    {
+      reass->last_heard = now;
+      return reass;
+    }
+
+  if (rm->reass_n >= rm->max_reass_n)
+    {
+      reass = NULL;
+      return reass;
+    }
+  else
+    {
+      pool_get (rm->pool, reass);
+      memset (reass, 0, sizeof (*reass));
+      reass->id = rm->id_counter;
+      ++rm->id_counter;
+      reass->first_bi = ~0;
+      reass->last_packet_octet = ~0;
+      reass->data_len = 0;
+      ++rm->reass_n;
+    }
+
+  reass->key.as_u64[0] = kv.key[0] = k->as_u64[0];
+  reass->key.as_u64[1] = kv.key[1] = k->as_u64[1];
+  reass->key.as_u64[2] = kv.key[2] = k->as_u64[2];
+  reass->key.as_u64[3] = kv.key[3] = k->as_u64[3];
+  reass->key.as_u64[4] = kv.key[4] = k->as_u64[4];
+  reass->key.as_u64[5] = kv.key[5] = k->as_u64[5];
+  kv.value = reass - rm->pool;
+  reass->last_heard = now;
+
+  if (clib_bihash_add_del_48_8 (&rm->hash, &kv, 1))
+    {
+      ip6_reass_free (rm, reass);
+      reass = NULL;
+    }
+
+  return reass;
+}
+
+void
+ip6_reass_finalize (vlib_main_t * vm, vlib_node_runtime_t * node,
+		    ip6_reass_main_t * rm, ip6_reass_t * reass, u32 * bi0,
+		    u32 * next0, vlib_error_t * error0, u32 next_input,
+		    u32 ** vec_drop_compress)
+{
+  ASSERT (~0 != reass->first_bi);
+  *bi0 = reass->first_bi;
+  *next0 = next_input;
+  *error0 = IP6_ERROR_NONE;
+  ip6_frag_hdr_t *frag_hdr;
+  vlib_buffer_t *last_b = NULL;
+  u32 sub_chain_bi = reass->first_bi;
+  u32 total_length = 0;
+  u32 buf_cnt = 0;
+  u32 dropped_cnt = 0;
+  do
+    {
+      u32 tmp_bi = sub_chain_bi;
+      vlib_buffer_t *tmp = vlib_get_buffer (vm, tmp_bi);
+      u32 data_len = ip6_reass_buffer_get_data_len (tmp);
+      u32 trim_front = vnet_buffer (tmp)->ip.reass.ip6_frag_hdr_offset +
+	sizeof (*frag_hdr) + ip6_reass_buffer_get_data_offset (tmp);
+      u32 trim_end =
+	vlib_buffer_length_in_chain (vm, tmp) - trim_front - data_len;
+      if (tmp_bi == reass->first_bi)
+	{
+	  /* first buffer - keep ip6 header */
+	  ASSERT (0 == ip6_reass_buffer_get_data_offset (tmp));
+	  trim_front = 0;
+	  trim_end = vlib_buffer_length_in_chain (vm, tmp) - data_len -
+	    (vnet_buffer (tmp)->ip.reass.ip6_frag_hdr_offset +
+	     sizeof (*frag_hdr));
+	}
+      u32 keep_data =
+	vlib_buffer_length_in_chain (vm, tmp) - trim_front - trim_end;
+      while (1)
+	{
+	  ++buf_cnt;
+	  if (trim_front)
+	    {
+	      if (trim_front > tmp->current_length)
+		{
+		  /* drop whole buffer */
+		  vec_add1 (*vec_drop_compress, tmp_bi);
+		  ++dropped_cnt;
+		  trim_front -= tmp->current_length;
+		  ASSERT (tmp->flags & VLIB_BUFFER_NEXT_PRESENT);
+		  tmp->flags &= ~VLIB_BUFFER_NEXT_PRESENT;
+		  tmp_bi = tmp->next_buffer;
+		  tmp = vlib_get_buffer (vm, tmp_bi);
+		  continue;
+		}
+	      else
+		{
+		  vlib_buffer_advance (tmp, trim_front);
+		  trim_front = 0;
+		}
+	    }
+	  if (keep_data)
+	    {
+	      if (last_b)
+		{
+		  last_b->flags |= VLIB_BUFFER_NEXT_PRESENT;
+		  last_b->next_buffer = tmp_bi;
+		}
+	      last_b = tmp;
+	      if (keep_data <= tmp->current_length)
+		{
+		  tmp->current_length = keep_data;
+		  keep_data = 0;
+		}
+	      else
+		{
+		  keep_data -= tmp->current_length;
+		  ASSERT (tmp->flags & VLIB_BUFFER_NEXT_PRESENT);
+		}
+	      total_length += tmp->current_length;
+	    }
+	  else
+	    {
+	      vec_add1 (*vec_drop_compress, tmp_bi);
+	      ASSERT (reass->first_bi != tmp_bi);
+	      ++dropped_cnt;
+	    }
+	  if (tmp->flags & VLIB_BUFFER_NEXT_PRESENT)
+	    {
+	      tmp_bi = tmp->next_buffer;
+	      tmp = vlib_get_buffer (vm, tmp->next_buffer);
+	    }
+	  else
+	    {
+	      break;
+	    }
+	}
+      sub_chain_bi =
+	vnet_buffer (vlib_get_buffer (vm, sub_chain_bi))->ip.
+	reass.next_range_bi;
+    }
+  while (~0 != sub_chain_bi);
+  last_b->flags &= ~VLIB_BUFFER_NEXT_PRESENT;
+  vlib_buffer_t *first_b = vlib_get_buffer (vm, reass->first_bi);
+  ASSERT (total_length >= first_b->current_length);
+  total_length -= first_b->current_length;
+  first_b->flags |= VLIB_BUFFER_TOTAL_LENGTH_VALID;
+  first_b->total_length_not_including_first_buffer = total_length;
+  // drop fragment header
+  vnet_buffer_opaque_t *first_b_vnb = vnet_buffer (first_b);
+  ip6_header_t *ip = vlib_buffer_get_current (first_b);
+  u16 ip6_frag_hdr_offset = first_b_vnb->ip.reass.ip6_frag_hdr_offset;
+  ip6_ext_header_t *prev_hdr;
+  ip6_ext_header_find_t (ip, prev_hdr, frag_hdr,
+			 IP_PROTOCOL_IPV6_FRAGMENTATION);
+  if (prev_hdr)
+    {
+      prev_hdr->next_hdr = frag_hdr->next_hdr;
+    }
+  else
+    {
+      ip->protocol = frag_hdr->next_hdr;
+    }
+  ASSERT ((u8 *) frag_hdr - (u8 *) ip == ip6_frag_hdr_offset);
+  memmove (frag_hdr, (u8 *) frag_hdr + sizeof (*frag_hdr),
+	   first_b->current_length - ip6_frag_hdr_offset -
+	   sizeof (ip6_frag_hdr_t));
+  first_b->current_length -= sizeof (*frag_hdr);
+  ip->payload_length =
+    clib_host_to_net_u16 (total_length + first_b->current_length -
+			  sizeof (*ip));
+  ip6_reass_free (rm, reass);
+  vlib_buffer_chain_compress (vm, first_b, vec_drop_compress);
+  if (PREDICT_FALSE (first_b->flags & VLIB_BUFFER_IS_TRACED))
+    {
+      ip6_reass_add_trace (vm, node, rm, reass, reass->first_bi, FINALIZE, 0);
+#if 0
+      // following code does a hexdump of packet fragments to stdout ...
+      do
+	{
+	  u32 bi = reass->first_bi;
+	  u8 *s = NULL;
+	  while (~0 != bi)
+	    {
+	      vlib_buffer_t *b = vlib_get_buffer (vm, bi);
+	      s = format (s, "%u: %U\n", bi, format_hexdump,
+			  vlib_buffer_get_current (b), b->current_length);
+	      if (b->flags & VLIB_BUFFER_NEXT_PRESENT)
+		{
+		  bi = b->next_buffer;
+		}
+	      else
+		{
+		  break;
+		}
+	    }
+	  printf ("%.*s\n", vec_len (s), s);
+	  fflush (stdout);
+	  vec_free (s);
+	}
+      while (0);
+#endif
+    }
+  reass = NULL;
+}
+
+static u32
+ip6_reass_get_buffer_chain_length (vlib_main_t * vm, vlib_buffer_t * b)
+{
+  u32 len = 0;
+  while (b)
+    {
+      ++len;
+      if (PREDICT_FALSE (b->flags & VLIB_BUFFER_NEXT_PRESENT))
+	{
+	  b = vlib_get_buffer (vm, b->next_buffer);
+	}
+      else
+	{
+	  break;
+	}
+    }
+  return len;
+}
+
+static void
+ip6_reass_insert_range_in_chain (vlib_main_t * vm,
+				 ip6_reass_main_t * rm,
+				 ip6_reass_t * reass,
+				 u32 prev_range_bi, u32 new_next_bi)
+{
+
+  vlib_buffer_t *new_next_b = vlib_get_buffer (vm, new_next_bi);
+  vnet_buffer_opaque_t *new_next_vnb = vnet_buffer (new_next_b);
+  if (~0 != prev_range_bi)
+    {
+      vlib_buffer_t *prev_b = vlib_get_buffer (vm, prev_range_bi);
+      vnet_buffer_opaque_t *prev_vnb = vnet_buffer (prev_b);
+      new_next_vnb->ip.reass.next_range_bi = prev_vnb->ip.reass.next_range_bi;
+      prev_vnb->ip.reass.next_range_bi = new_next_bi;
+    }
+  else
+    {
+      if (~0 != reass->first_bi)
+	{
+	  new_next_vnb->ip.reass.next_range_bi = reass->first_bi;
+	}
+      reass->first_bi = new_next_bi;
+    }
+  reass->data_len += ip6_reass_buffer_get_data_len (new_next_b);
+  rm->buffers_n += ip6_reass_get_buffer_chain_length (vm, new_next_b);
+}
+
+void
+ip6_reass_update (vlib_main_t * vm, vlib_node_runtime_t * node,
+		  ip6_reass_main_t * rm, ip6_reass_t * reass, u32 * bi0,
+		  u32 * next0, vlib_error_t * error0,
+		  ip6_frag_hdr_t * frag_hdr, u32 ** vec_drop_overlap,
+		  u32 ** vec_drop_compress, u32 next_input, u32 next_drop,
+		  u32 next_icmp_error)
+{
+  int consumed = 0;
+  vlib_buffer_t *fb = vlib_get_buffer (vm, *bi0);
+  vnet_buffer_opaque_t *fvnb = vnet_buffer (fb);
+  fvnb->ip.reass.ip6_frag_hdr_offset =
+    (u8 *) frag_hdr - (u8 *) vlib_buffer_get_current (fb);
+  ip6_header_t *fip = vlib_buffer_get_current (fb);
+  ASSERT (fb->current_length > sizeof (*fip));
+  ASSERT (fvnb->ip.reass.ip6_frag_hdr_offset > 0 &&
+	  fvnb->ip.reass.ip6_frag_hdr_offset < fb->current_length);
+  u32 fragment_first = fvnb->ip.reass.fragment_first =
+    ip6_frag_hdr_offset_bytes (frag_hdr);
+  u32 fragment_length =
+    vlib_buffer_length_in_chain (vm, fb) -
+    (fvnb->ip.reass.ip6_frag_hdr_offset + sizeof (*frag_hdr));
+  u32 fragment_last = fvnb->ip.reass.fragment_last =
+    fragment_first + fragment_length - 1;
+  int more_fragments = ip6_frag_hdr_more (frag_hdr);
+  u32 candidate_range_bi = reass->first_bi;
+  u32 prev_range_bi = ~0;
+  fvnb->ip.reass.range_first = fragment_first;
+  fvnb->ip.reass.range_last = fragment_last;
+  fvnb->ip.reass.next_range_bi = ~0;
+  if (more_fragments && 0 != fragment_length % 8)
+    {
+      if (PREDICT_FALSE (fb->flags & VLIB_BUFFER_IS_TRACED))
+	{
+	  ip6_reass_add_trace (vm, node, rm, reass, *bi0,
+			       ICMP_ERROR_FL_NOT_MULT_8, 0);
+	}
+      *next0 = next_icmp_error;
+      icmp6_error_set_vnet_buffer (fb, ICMP6_parameter_problem,
+				   ICMP6_parameter_problem_erroneous_header_field,
+				   (u8 *) & fip->payload_length - (u8 *) fip);
+      return;
+    }
+  if (fragment_first + fragment_length > 65535)
+    {
+      if (PREDICT_FALSE (fb->flags & VLIB_BUFFER_IS_TRACED))
+	{
+	  ip6_reass_add_trace (vm, node, rm, reass, *bi0,
+			       ICMP_ERROR_FL_TOO_BIG, 0);
+	}
+      *next0 = next_icmp_error;
+      ip6_header_t *ip0 = vlib_buffer_get_current (fb);
+      icmp6_error_set_vnet_buffer (fb, ICMP6_parameter_problem,
+				   ICMP6_parameter_problem_erroneous_header_field,
+				   (u8 *) & frag_hdr->fragment_offset_and_more
+				   - (u8 *) ip0);
+      return;
+    }
+  if (!more_fragments)
+    {
+      reass->last_packet_octet = fragment_last;
+    }
+  if (~0 == reass->first_bi)
+    {
+      // starting a new reassembly
+      ip6_reass_insert_range_in_chain (vm, rm, reass, prev_range_bi, *bi0);
+      if (PREDICT_FALSE (fb->flags & VLIB_BUFFER_IS_TRACED))
+	{
+	  ip6_reass_add_trace (vm, node, rm, reass, *bi0, RANGE_NEW, 0);
+	}
+      *bi0 = ~0;
+      return;
+    }
+  fvnb->ip.reass.estimated_mtu =
+    clib_min (clib_net_to_host_u16 (fip->payload_length),
+	      fvnb->ip.reass.estimated_mtu);
+  while (~0 != candidate_range_bi)
+    {
+      vlib_buffer_t *candidate_b = vlib_get_buffer (vm, candidate_range_bi);
+      vnet_buffer_opaque_t *candidate_vnb = vnet_buffer (candidate_b);
+      if (fragment_first > candidate_vnb->ip.reass.range_last)
+	{
+	  // this fragments starts after candidate range
+	  prev_range_bi = candidate_range_bi;
+	  candidate_range_bi = candidate_vnb->ip.reass.next_range_bi;
+	  if (candidate_vnb->ip.reass.range_last < fragment_last &&
+	      ~0 == candidate_range_bi)
+	    {
+	      // special case - this fragment falls beyond all known ranges
+	      ip6_reass_insert_range_in_chain (vm, rm, reass, prev_range_bi,
+					       *bi0);
+	      consumed = 1;
+	      break;
+	    }
+	  continue;
+	}
+      if (fragment_last < candidate_vnb->ip.reass.range_first)
+	{
+	  // this fragment ends before candidate range without any overlap
+	  ip6_reass_insert_range_in_chain (vm, rm, reass, prev_range_bi,
+					   *bi0);
+	  consumed = 1;
+	}
+      else if (fragment_first == candidate_vnb->ip.reass.range_first &&
+	       fragment_last == candidate_vnb->ip.reass.range_last)
+	{
+	  // duplicate fragment - ignore
+	}
+      else
+	{
+	  // overlapping fragment - not allowed by RFC 8200
+	  ip6_reass_drop_all (vm, rm, reass, vec_drop_overlap);
+	  ip6_reass_free (rm, reass);
+	  if (PREDICT_FALSE (fb->flags & VLIB_BUFFER_IS_TRACED))
+	    {
+	      ip6_reass_add_trace (vm, node, rm, reass, *bi0, RANGE_OVERLAP,
+				   0);
+	    }
+	  *next0 = next_drop;
+	  *error0 = IP6_ERROR_REASS_OVERLAPPING_FRAGMENT;
+	}
+      break;
+    }
+  if (consumed)
+    {
+      if (PREDICT_FALSE (fb->flags & VLIB_BUFFER_IS_TRACED))
+	{
+	  ip6_reass_add_trace (vm, node, rm, reass, *bi0, RANGE_NEW, 0);
+	}
+    }
+  if (~0 != reass->last_packet_octet &&
+      reass->data_len == reass->last_packet_octet + 1)
+    {
+      ip6_reass_finalize (vm, node, rm, reass, bi0, next0, error0, next_input,
+			  vec_drop_compress);
+    }
+  else
+    {
+      if (consumed)
+	{
+	  *bi0 = ~0;
+	}
+      else
+	{
+	  *next0 = next_drop;
+	  *error0 = IP6_ERROR_REASS_DUPLICATE_FRAGMENT;
+	}
+    }
+}
+
+always_inline uword
+ip6_reassembly (vlib_main_t * vm, vlib_node_runtime_t * node,
+		vlib_frame_t * frame)
+{
+  u32 *from = vlib_frame_vector_args (frame);
+  u32 n_left_from, n_left_to_next, *to_next, next_index;
+  ip6_reass_main_t *rm = &ip6_reass_main;
+
+  n_left_from = frame->n_vectors;
+  next_index = node->cached_next_index;
+  static u32 *vec_timeout = NULL;	// indexes of buffers which timed out
+  static u32 *vec_drop_overlap = NULL;	// indexes of buffers dropped due to overlap
+  static u32 *vec_drop_compress = NULL;	// indexes of buffers dropped due to buffer compression
+  while (n_left_from > 0 || vec_len (vec_timeout) > 0 ||
+	 vec_len (vec_drop_overlap) > 0)
+    {
+      vlib_get_next_frame (vm, node, next_index, to_next, n_left_to_next);
+
+      while (vec_len (vec_timeout) > 0 && n_left_to_next > 0)
+	{
+	  u32 bi = vec_pop (vec_timeout);
+	  vlib_buffer_t *b = vlib_get_buffer (vm, bi);
+	  b->error = node->errors[IP6_ERROR_REASS_TIMEOUT];
+	  to_next[0] = bi;
+	  to_next += 1;
+	  n_left_to_next -= 1;
+	  vlib_validate_buffer_enqueue_x1 (vm, node, next_index, to_next,
+					   n_left_to_next, bi,
+					   IP6_REASSEMBLY_NEXT_DROP);
+	  ASSERT (rm->buffers_n > 0);
+	  --rm->buffers_n;
+	}
+
+      while (vec_len (vec_drop_overlap) > 0 && n_left_to_next > 0)
+	{
+	  u32 bi = vec_pop (vec_drop_overlap);
+	  vlib_buffer_t *b = vlib_get_buffer (vm, bi);
+	  b->error = node->errors[IP6_ERROR_REASS_OVERLAPPING_FRAGMENT];
+	  to_next[0] = bi;
+	  to_next += 1;
+	  n_left_to_next -= 1;
+	  vlib_validate_buffer_enqueue_x1 (vm, node, next_index, to_next,
+					   n_left_to_next, bi,
+					   IP6_REASSEMBLY_NEXT_DROP);
+	  ASSERT (rm->buffers_n > 0);
+	  --rm->buffers_n;
+	}
+
+      while (vec_len (vec_drop_compress) > 0 && n_left_to_next > 0)
+	{
+	  u32 bi = vec_pop (vec_drop_compress);
+	  vlib_buffer_t *b = vlib_get_buffer (vm, bi);
+	  b->error = node->errors[IP6_ERROR_NONE];
+	  to_next[0] = bi;
+	  to_next += 1;
+	  n_left_to_next -= 1;
+	  vlib_validate_buffer_enqueue_x1 (vm, node, next_index, to_next,
+					   n_left_to_next, bi,
+					   IP6_REASSEMBLY_NEXT_DROP);
+	  ASSERT (rm->buffers_n > 0);
+	  --rm->buffers_n;
+	}
+
+      while (n_left_from > 0 && n_left_to_next > 0)
+	{
+	  u32 bi0;
+	  vlib_buffer_t *b0;
+	  u32 next0;		//, error0;
+
+	  bi0 = from[0];
+	  b0 = vlib_get_buffer (vm, bi0);
+
+	  ip6_header_t *ip0 = vlib_buffer_get_current (b0);
+	  ip6_frag_hdr_t *frag_hdr;
+	  ip6_ext_header_t *prev_hdr;
+	  ip6_ext_header_find_t (ip0, prev_hdr, frag_hdr,
+				 IP_PROTOCOL_IPV6_FRAGMENTATION);
+	  if (0 == ip6_frag_hdr_offset (frag_hdr))
+	    {
+	      // first fragment - verify upper-layer is present
+	      ip6_ext_header_t *tmp = (ip6_ext_header_t *) frag_hdr;
+	      while (ip6_ext_hdr (tmp->next_hdr))
+		{
+		  tmp = ip6_ext_next_header (tmp);
+		}
+	      if (IP_PROTOCOL_IP6_NONXT == tmp->next_hdr)
+		{
+		  icmp6_error_set_vnet_buffer (b0, ICMP6_parameter_problem,
+					       ICMP6_parameter_problem_first_fragment_has_incomplete_header_chain,
+					       0);
+		  b0->error = node->errors[IP6_ERROR_REASS_MISSING_UPPER];
+
+		  to_next[0] = bi0;
+		  to_next += 1;
+		  n_left_to_next -= 1;
+		  vlib_validate_buffer_enqueue_x1 (vm, node, next_index,
+						   to_next, n_left_to_next,
+						   bi0,
+						   IP6_REASSEMBLY_NEXT_ICMP_ERROR);
+		  goto next;
+		}
+	    }
+	  vnet_buffer (b0)->ip.reass.ip6_frag_hdr_offset =
+	    (u8 *) frag_hdr - (u8 *) ip0;
+
+	  ip6_reass_key_t k;
+	  k.src.as_u64[0] = ip0->src_address.as_u64[0];
+	  k.src.as_u64[1] = ip0->src_address.as_u64[1];
+	  k.dst.as_u64[0] = ip0->dst_address.as_u64[0];
+	  k.dst.as_u64[1] = ip0->dst_address.as_u64[1];
+	  k.xx_id = vnet_buffer (b0)->sw_if_index[VLIB_RX];
+	  k.frag_id = frag_hdr->identification;
+	  k.proto = ip0->protocol;
+	  k.unused = 0;
+	  u32 icmp_bi = ~0;
+	  ip6_reass_t *reass =
+	    ip6_reass_find_or_create (vm, node, rm, &k, &icmp_bi,
+				      &vec_timeout);
+
+	  u32 error0 = IP6_ERROR_NONE;
+	  if (reass)
+	    {
+	      ip6_reass_update (vm, node, rm, reass, &bi0, &next0, &error0,
+				frag_hdr, &vec_drop_overlap,
+				&vec_drop_compress, IP6_REASSEMBLY_NEXT_INPUT,
+				IP6_REASSEMBLY_NEXT_DROP,
+				IP6_REASSEMBLY_NEXT_ICMP_ERROR);
+	    }
+	  else
+	    {
+	      next0 = IP6_REASSEMBLY_NEXT_DROP;
+	      error0 = IP6_ERROR_REASS_LIMIT_REACHED;
+	    }
+
+	  b0->error = node->errors[error0];
+
+	  if (~0 != bi0)
+	    {
+	      to_next[0] = bi0;
+	      to_next += 1;
+	      n_left_to_next -= 1;
+	      vlib_validate_buffer_enqueue_x1 (vm, node, next_index, to_next,
+					       n_left_to_next, bi0, next0);
+	    }
+
+	  if (~0 != icmp_bi)
+	    {
+	      next0 = IP6_REASSEMBLY_NEXT_ICMP_ERROR;
+	      to_next[0] = icmp_bi;
+	      to_next += 1;
+	      n_left_to_next -= 1;
+	      vlib_validate_buffer_enqueue_x1 (vm, node, next_index, to_next,
+					       n_left_to_next, icmp_bi,
+					       next0);
+	    }
+	next:
+	  from += 1;
+	  n_left_from -= 1;
+	}
+
+      vlib_put_next_frame (vm, node, next_index, n_left_to_next);
+    }
+
+  return frame->n_vectors;
+}
+
+static char *ip6_reassembly_error_strings[] = {
+#define _(sym, string) string,
+  foreach_ip6_error
+#undef _
+};
+
+/* *INDENT-OFF* */
+VLIB_REGISTER_NODE (ip6_reass_node, static) = {
+    .function = ip6_reassembly,
+    .name = "ip6-reassembly",
+    .vector_size = sizeof (u32),
+    .format_trace = format_ip6_reass_trace,
+    .n_errors = ARRAY_LEN (ip6_reassembly_error_strings),
+    .error_strings = ip6_reassembly_error_strings,
+    .n_next_nodes = IP6_REASSEMBLY_N_NEXT,
+    .next_nodes =
+        {
+                [IP6_REASSEMBLY_NEXT_INPUT] = "ip6-input",
+                [IP6_REASSEMBLY_NEXT_DROP] = "ip6-drop",
+                [IP6_REASSEMBLY_NEXT_ICMP_ERROR] = "ip6-icmp-error",
+        },
+};
+/* *INDENT-ON* */
+
+VLIB_NODE_FUNCTION_MULTIARCH (ip6_reass_node, ip6_reassembly)
+     static u32 ip6_reass_get_nbuckets ()
+{
+  ip6_reass_main_t *rm = &ip6_reass_main;
+  u32 nbuckets;
+  u8 i;
+
+  nbuckets = (u32) (rm->max_reass_n / IP6_REASS_HT_LOAD_FACTOR);
+
+  for (i = 0; i < 31; i++)
+    if ((1 << i) >= nbuckets)
+      break;
+  nbuckets = 1 << i;
+
+  return nbuckets;
+}
+
+typedef enum
+{
+  IP6_EVENT_CONFIG_CHANGED = 1,
+} ip6_reass_event_t;
+
+typedef struct
+{
+  int failure;
+  clib_bihash_48_8_t *new_hash;
+} ip6_rehash_cb_ctx;
+
+void
+ip6_rehash_cb (clib_bihash_kv_48_8_t * kv, void *_ctx)
+{
+  ip6_rehash_cb_ctx *ctx = _ctx;
+  if (clib_bihash_add_del_48_8 (ctx->new_hash, kv, 1))
+    {
+      ctx->failure = 1;
+    }
+}
+
+vnet_api_error_t
+ip6_reass_set (u32 timeout_ms, u32 max_reassemblies,
+	       u32 expire_walk_interval_ms)
+{
+  u32 old_nbuckets = ip6_reass_get_nbuckets ();
+  ip6_reass_main.timeout_ms = timeout_ms;
+  ip6_reass_main.timeout = (f64) timeout_ms / (f64) MSEC_PER_SEC;
+  ip6_reass_main.max_reass_n = max_reassemblies;
+  ip6_reass_main.expire_walk_interval_ms = expire_walk_interval_ms;
+  vlib_process_signal_event (ip6_reass_main.vlib_main,
+			     ip6_reass_main.ip6_reass_expire_node_idx,
+			     IP6_EVENT_CONFIG_CHANGED, 0);
+  u32 new_nbuckets = ip6_reass_get_nbuckets ();
+  if (ip6_reass_main.max_reass_n > 0 && new_nbuckets > 1 &&
+      new_nbuckets != old_nbuckets)
+    {
+      clib_bihash_48_8_t new_hash;
+      memset (&new_hash, 0, sizeof (new_hash));
+      ip6_rehash_cb_ctx ctx;
+      ctx.failure = 0;
+      ctx.new_hash = &new_hash;
+      clib_bihash_init_48_8 (&new_hash, "ip6-reass", new_nbuckets,
+			     new_nbuckets * 1024);
+      clib_bihash_foreach_key_value_pair_48_8 (&ip6_reass_main.hash,
+					       ip6_rehash_cb, &ctx);
+      if (ctx.failure)
+	{
+	  clib_bihash_free_48_8 (&new_hash);
+	  return -1;
+	}
+      else
+	{
+	  clib_bihash_free_48_8 (&ip6_reass_main.hash);
+	  clib_memcpy (&ip6_reass_main.hash, &new_hash,
+		       sizeof (ip6_reass_main.hash));
+	}
+    }
+  return 0;
+}
+
+vnet_api_error_t
+ip6_reass_get (u32 * timeout_ms, u32 * max_reassemblies,
+	       u32 * expire_walk_interval_ms)
+{
+  *timeout_ms = ip6_reass_main.timeout_ms;
+  *max_reassemblies = ip6_reass_main.max_reass_n;
+  *expire_walk_interval_ms = ip6_reass_main.expire_walk_interval_ms;
+  return 0;
+}
+
+clib_error_t *
+ip6_reass_init_function (vlib_main_t * vm)
+{
+  ip6_reass_main_t *rm = &ip6_reass_main;
+  clib_error_t *error = 0;
+  u32 nbuckets;
+
+  rm->vlib_main = vm;
+  rm->vnet_main = vnet_get_main ();
+
+  rm->reass_n = 0;
+  pool_alloc (rm->pool, rm->max_reass_n);
+  ip6_reass_set (IP6_REASS_TIMEOUT_DEFAULT_MS,
+		 IP6_REASS_MAX_REASSEMBLIES_DEAFULT,
+		 IP6_REASS_EXPIRE_WALK_INTERVAL_DEFAULT_MS);
+
+  nbuckets = ip6_reass_get_nbuckets ();
+  clib_bihash_init_48_8 (&rm->hash, "ip6-reass", nbuckets, nbuckets * 1024);
+
+  vlib_node_t *node = vlib_get_node_by_name (vm, (u8 *) "ip6-drop");
+  ASSERT (node);
+  rm->ip6_drop_idx = node->index;
+  node = vlib_get_node_by_name (vm, (u8 *) "ip6-reassembly-expire-walk");
+  ASSERT (node);
+  rm->ip6_reass_expire_node_idx = node->index;
+  node = vlib_get_node_by_name (vm, (u8 *) "ip6-icmp-error");
+  ASSERT (node);
+  rm->ip6_icmp_error_idx = node->index;
+
+  if ((error = vlib_call_init_function (vm, ip_main_init)))
+    return error;
+  ip6_register_protocol (IP_PROTOCOL_IPV6_FRAGMENTATION,
+			 ip6_reass_node.index);
+  return error;
+}
+
+VLIB_INIT_FUNCTION (ip6_reass_init_function);
+
+static uword
+ip6_reass_walk_expired (vlib_main_t * vm,
+			vlib_node_runtime_t * node, vlib_frame_t * f)
+{
+  ip6_reass_main_t *rm = &ip6_reass_main;
+  uword event_type, *event_data = 0;
+
+  while (true)
+    {
+      vlib_process_wait_for_event_or_clock (vm,
+					    (f64) rm->expire_walk_interval_ms
+					    / (f64) MSEC_PER_SEC);
+      event_type = vlib_process_get_events (vm, &event_data);
+
+      switch (event_type)
+	{
+	case ~0:		/* no events => timeout */
+	  /* nothing to do here */
+	  break;
+	case IP6_EVENT_CONFIG_CHANGED:
+	  break;
+	default:
+	  clib_warning ("BUG: event type 0x%wx", event_type);
+	  break;
+	}
+      f64 now = vlib_time_now (vm);
+
+      ip6_reass_t *reass;
+      u32 *vec_timeout = NULL;
+      int *pool_indexes_to_free = NULL;
+
+      int index;
+      /* *INDENT-OFF* */
+      pool_foreach_index (index, rm->pool, ({
+                            reass = pool_elt_at_index (rm->pool, index);
+                            if (now > reass->last_heard + rm->timeout)
+                              {
+                                vec_add1 (pool_indexes_to_free, index);
+                              }
+                          }));
+      /* *INDENT-ON* */
+      int *i;
+      u32 *vec_icmp_bi = NULL;
+      /* *INDENT-OFF* */
+      vec_foreach (i, pool_indexes_to_free)
+      {
+        ip6_reass_t *reass = pool_elt_at_index (rm->pool, i[0]);
+        u32 icmp_bi = ~0;
+        ip6_reass_on_timeout (vm, node, rm, reass, &icmp_bi, &vec_timeout);
+        if (~0 != icmp_bi)
+          {
+            vec_add1 (vec_icmp_bi, icmp_bi);
+          }
+        ip6_reass_free (rm, reass);
+      }
+      /* *INDENT-ON* */
+
+      while (vec_len (vec_timeout) > 0)
+	{
+	  vlib_frame_t *f = vlib_get_frame_to_node (vm, rm->ip6_drop_idx);
+	  u32 *to_next = vlib_frame_vector_args (f);
+	  u32 n_left_to_next = VLIB_FRAME_SIZE - f->n_vectors;
+	  u32 n_trace = 0;
+	  while (vec_len (vec_timeout) > 0 && n_left_to_next > 0)
+	    {
+	      u32 bi = vec_pop (vec_timeout);
+	      vlib_buffer_t *b = vlib_get_buffer (vm, bi);
+	      if (PREDICT_FALSE (b->flags & VLIB_BUFFER_IS_TRACED))
+		{
+		  if (pool_is_free_index (vm->trace_main.trace_buffer_pool,
+					  b->trace_index))
+		    {
+		      /* the trace is gone, don't trace this buffer anymore */
+		      b->flags &= ~VLIB_BUFFER_IS_TRACED;
+		    }
+		  else
+		    {
+		      ++n_trace;
+		    }
+		}
+	      b->error = node->errors[IP6_ERROR_REASS_TIMEOUT];
+	      to_next[0] = bi;
+	      ++f->n_vectors;
+	      to_next += 1;
+	      n_left_to_next -= 1;
+	      ASSERT (rm->buffers_n > 0);
+	      --rm->buffers_n;
+	    }
+	  if (PREDICT_FALSE (n_trace > 0))
+	    {
+	      f->flags |= VLIB_FRAME_TRACE;
+	    }
+	  vlib_put_frame_to_node (vm, rm->ip6_drop_idx, f);
+	}
+
+      while (vec_len (vec_icmp_bi) > 0)
+	{
+	  vlib_frame_t *f =
+	    vlib_get_frame_to_node (vm, rm->ip6_icmp_error_idx);
+	  u32 *to_next = vlib_frame_vector_args (f);
+	  u32 n_left_to_next = VLIB_FRAME_SIZE - f->n_vectors;
+	  u32 n_trace = 0;
+	  while (vec_len (vec_icmp_bi) > 0 && n_left_to_next > 0)
+	    {
+	      u32 bi = vec_pop (vec_icmp_bi);
+	      vlib_buffer_t *b = vlib_get_buffer (vm, bi);
+	      if (PREDICT_FALSE (b->flags & VLIB_BUFFER_IS_TRACED))
+		{
+		  if (pool_is_free_index (vm->trace_main.trace_buffer_pool,
+					  b->trace_index))
+		    {
+		      /* the trace is gone, don't trace this buffer anymore */
+		      b->flags &= ~VLIB_BUFFER_IS_TRACED;
+		    }
+		  else
+		    {
+		      ++n_trace;
+		    }
+		}
+	      b->error = node->errors[IP6_ERROR_REASS_TIMEOUT];
+	      to_next[0] = bi;
+	      ++f->n_vectors;
+	      to_next += 1;
+	      n_left_to_next -= 1;
+	      ASSERT (rm->buffers_n > 0);
+	      --rm->buffers_n;
+	    }
+	  if (PREDICT_FALSE (n_trace > 0))
+	    {
+	      f->flags |= VLIB_FRAME_TRACE;
+	    }
+	  vlib_put_frame_to_node (vm, rm->ip6_icmp_error_idx, f);
+	}
+
+      vec_free (pool_indexes_to_free);
+      vec_free (vec_timeout);
+      vec_free (vec_icmp_bi);
+      if (event_data)
+	{
+	  _vec_len (event_data) = 0;
+	}
+    }
+
+  return 0;
+}
+
+static vlib_node_registration_t ip6_reass_expire_node;
+
+/* *INDENT-OFF* */
+VLIB_REGISTER_NODE (ip6_reass_expire_node, static) = {
+    .function = ip6_reass_walk_expired,
+    .format_trace = format_ip6_reass_trace,
+    .type = VLIB_NODE_TYPE_PROCESS,
+    .name = "ip6-reassembly-expire-walk",
+
+    .n_errors = ARRAY_LEN (ip6_reassembly_error_strings),
+    .error_strings = ip6_reassembly_error_strings,
+
+};
+/* *INDENT-ON* */
+
+static u8 *
+format_ip6_reass_key (u8 * s, va_list * args)
+{
+  ip6_reass_key_t *key = va_arg (*args, ip6_reass_key_t *);
+  s = format (s, "xx_id: %u, src: %U, dst: %U, frag_id: %u, proto: %u",
+	      key->xx_id, format_ip6_address, &key->src, format_ip6_address,
+	      &key->dst, clib_net_to_host_u16 (key->frag_id), key->proto);
+  return s;
+}
+
+static u8 *
+format_ip6_reass (u8 * s, va_list * args)
+{
+  vlib_main_t *vm = va_arg (*args, vlib_main_t *);
+  ip6_reass_t *reass = va_arg (*args, ip6_reass_t *);
+
+  s = format (s, "ID: %u, key: %U\n  first_bi: %u, data_len: %u, "
+	      "last_packet_octet: %u, trace_op_counter: %u\n",
+	      reass->id, format_ip6_reass_key, &reass->key, reass->first_bi,
+	      reass->data_len, reass->last_packet_octet,
+	      reass->trace_op_counter);
+  u32 bi = reass->first_bi;
+  u32 counter = 0;
+  while (~0 != bi)
+    {
+      vlib_buffer_t *b = vlib_get_buffer (vm, bi);
+      vnet_buffer_opaque_t *vnb = vnet_buffer (b);
+      s = format (s, "  #%03u: range: [%u, %u], bi: %u, off: %d, len: %u, "
+		  "fragment[%u, %u]\n",
+		  counter, vnb->ip.reass.range_first,
+		  vnb->ip.reass.range_last, bi,
+		  ip6_reass_buffer_get_data_offset_no_check (b),
+		  ip6_reass_buffer_get_data_len_no_check (b),
+		  vnb->ip.reass.fragment_first, vnb->ip.reass.fragment_last);
+      if (b->flags & VLIB_BUFFER_NEXT_PRESENT)
+	{
+	  bi = b->next_buffer;
+	}
+      else
+	{
+	  bi = ~0;
+	}
+    }
+  return s;
+}
+
+static clib_error_t *
+show_ip6_reass (vlib_main_t * vm, unformat_input_t * input,
+		CLIB_UNUSED (vlib_cli_command_t * lmd))
+{
+  ip6_reass_main_t *rm = &ip6_reass_main;
+
+  vlib_cli_output (vm, "---------------------");
+  vlib_cli_output (vm, "IP6 reassembly status");
+  vlib_cli_output (vm, "---------------------");
+  if (unformat (input, "details"))
+    {
+      ip6_reass_t *reass;
+      /* *INDENT-OFF* */
+      pool_foreach (reass, rm->pool, {
+        vlib_cli_output (vm, "%U", format_ip6_reass, vm, reass);
+      });
+      /* *INDENT-ON* */
+    }
+  vlib_cli_output (vm, "---------------------");
+  vlib_cli_output (vm, "Current IP6 reassemblies count: %lu\n", rm->reass_n);
+  vlib_cli_output (vm,
+		   "Maximum configured concurrent IP6 reassemblies: %lu\n",
+		   (long unsigned) rm->max_reass_n);
+  vlib_cli_output (vm, "Buffers in use: %lu\n",
+		   (long unsigned) rm->buffers_n);
+  return 0;
+}
+
+/* *INDENT-OFF* */
+VLIB_CLI_COMMAND (show_ip6_reassembly_cmd, static) = {
+    .path = "show ip6-reassembly",
+    .short_help = "show ip6-reassembly [details]",
+    .function = show_ip6_reass,
+};
+/* *INDENT-ON* */
+
+/*
+ * fd.io coding-style-patch-verification: ON
+ *
+ * Local Variables:
+ * eval: (c-set-style "gnu")
+ * End:
+ */
diff --git a/src/vnet/ip/ip6_reassembly.h b/src/vnet/ip/ip6_reassembly.h
new file mode 100644
index 00000000000..0de4e04bb2b
--- /dev/null
+++ b/src/vnet/ip/ip6_reassembly.h
@@ -0,0 +1,49 @@
+/*
+ * Copyright (c) 2017 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 IPv6 Reassembly.
+ *
+ * This file contains the source code for IPv6 reassembly.
+ */
+
+#ifndef __included_ip6_reassembly_h__
+#define __included_ip6_reassembly_h__
+
+#include <vnet/api_errno.h>
+#include <vnet/vnet.h>
+
+/**
+ * @brief set ip6 reassembly configuration
+ */
+vnet_api_error_t ip6_reass_set (u32 timeout_ms, u32 max_reassemblies,
+				u32 expire_walk_interval_ms);
+
+/**
+ * @brief get ip6 reassembly configuration
+ */
+vnet_api_error_t ip6_reass_get (u32 * timeout_ms, u32 * max_reassemblies,
+				u32 * expire_walk_interval_ms);
+
+#endif /* __included_ip6_reassembly_h */
+
+/*
+ * fd.io coding-style-patch-verification: ON
+ *
+ * Local Variables:
+ * eval: (c-set-style "gnu")
+ * End:
+ */
diff --git a/src/vnet/ip/ip_api.c b/src/vnet/ip/ip_api.c
index c0889ebfa6b..60fa2faf926 100644
--- a/src/vnet/ip/ip_api.c
+++ b/src/vnet/ip/ip_api.c
@@ -42,6 +42,8 @@
 #include <vnet/fib/ip4_fib.h>
 #include <vnet/fib/ip6_fib.h>
 #include <vnet/ip/ip6_hop_by_hop.h>
+#include <vnet/ip/ip4_reassembly.h>
+#include <vnet/ip/ip6_reassembly.h>
 
 #include <vnet/vnet_msg_enum.h>
 
@@ -97,7 +99,9 @@ _(IOAM_DISABLE, ioam_disable)                                           \
 _(IP_SOURCE_AND_PORT_RANGE_CHECK_ADD_DEL,                               \
   ip_source_and_port_range_check_add_del)                               \
 _(IP_SOURCE_AND_PORT_RANGE_CHECK_INTERFACE_ADD_DEL,                     \
-  ip_source_and_port_range_check_interface_add_del)
+  ip_source_and_port_range_check_interface_add_del)                     \
+_(IP_REASSEMBLY_SET, ip_reassembly_set)                                 \
+_(IP_REASSEMBLY_GET, ip_reassembly_get)
 
 extern void stats_dslock_with_hint (int hint, int tag);
 extern void stats_dsunlock (void);
@@ -2786,6 +2790,61 @@ vl_api_set_arp_neighbor_limit_t_handler (vl_api_set_arp_neighbor_limit_t * mp)
   REPLY_MACRO (VL_API_SET_ARP_NEIGHBOR_LIMIT_REPLY);
 }
 
+void
+vl_api_ip_reassembly_set_t_handler (vl_api_ip_reassembly_set_t * mp)
+{
+  vl_api_ip_reassembly_set_reply_t *rmp;
+  int rv = 0;
+  if (mp->is_ip6)
+    {
+      rv = ip6_reass_set (clib_net_to_host_u32 (mp->timeout_ms),
+			  clib_net_to_host_u32 (mp->max_reassemblies),
+			  clib_net_to_host_u32 (mp->expire_walk_interval_ms));
+    }
+  else
+    {
+      rv = ip4_reass_set (clib_net_to_host_u32 (mp->timeout_ms),
+			  clib_net_to_host_u32 (mp->max_reassemblies),
+			  clib_net_to_host_u32 (mp->expire_walk_interval_ms));
+    }
+
+  REPLY_MACRO (VL_API_IP_REASSEMBLY_SET_REPLY);
+}
+
+void
+vl_api_ip_reassembly_get_t_handler (vl_api_ip_reassembly_get_t * mp)
+{
+  unix_shared_memory_queue_t *q;
+
+  q = vl_api_client_index_to_input_queue (mp->client_index);
+
+  if (q == 0)
+    return;
+
+  vl_api_ip_reassembly_get_reply_t *rmp = vl_msg_api_alloc (sizeof (*rmp));
+  memset (rmp, 0, sizeof (*rmp));
+  rmp->_vl_msg_id = ntohs (VL_API_IP_REASSEMBLY_GET_REPLY);
+  rmp->context = mp->context;
+  rmp->retval = 0;
+  if (mp->is_ip6)
+    {
+      rmp->is_ip6 = 1;
+      ip6_reass_get (&rmp->timeout_ms, &rmp->max_reassemblies,
+		     &rmp->expire_walk_interval_ms);
+    }
+  else
+    {
+      rmp->is_ip6 = 0;
+      ip4_reass_get (&rmp->timeout_ms, &rmp->max_reassemblies,
+		     &rmp->expire_walk_interval_ms);
+    }
+  rmp->timeout_ms = clib_host_to_net_u32 (rmp->timeout_ms);
+  rmp->max_reassemblies = clib_host_to_net_u32 (rmp->max_reassemblies);
+  rmp->expire_walk_interval_ms =
+    clib_host_to_net_u32 (rmp->expire_walk_interval_ms);
+  vl_msg_api_send_shmem (q, (u8 *) & rmp);
+}
+
 #define vl_msg_name_crc_list
 #include <vnet/ip/ip.api.h>
 #undef vl_msg_name_crc_list
diff --git a/src/vnet/ip/lookup.c b/src/vnet/ip/lookup.c
index 67e46bd8cfa..f2880bf8cd9 100644
--- a/src/vnet/ip/lookup.c
+++ b/src/vnet/ip/lookup.c
@@ -230,6 +230,8 @@ ip_lookup_init (ip_lookup_main_t * lm, u32 is_ip6)
       }
 
     lm->local_next_by_ip_protocol[IP_PROTOCOL_UDP] = IP_LOCAL_NEXT_UDP_LOOKUP;
+    lm->local_next_by_ip_protocol[IP_PROTOCOL_VPP_FRAGMENTATION] =
+      IP_LOCAL_NEXT_REASSEMBLY;
     lm->local_next_by_ip_protocol[is_ip6 ? IP_PROTOCOL_ICMP6 :
 				  IP_PROTOCOL_ICMP] = IP_LOCAL_NEXT_ICMP;
     lm->builtin_protocol_by_ip_protocol[IP_PROTOCOL_UDP] =
diff --git a/src/vnet/ip/lookup.h b/src/vnet/ip/lookup.h
index a89546f5d92..9fe7a1b0085 100644
--- a/src/vnet/ip/lookup.h
+++ b/src/vnet/ip/lookup.h
@@ -111,6 +111,7 @@ typedef enum
   IP_LOCAL_NEXT_PUNT,
   IP_LOCAL_NEXT_UDP_LOOKUP,
   IP_LOCAL_NEXT_ICMP,
+  IP_LOCAL_NEXT_REASSEMBLY,
   IP_LOCAL_N_NEXT,
 } ip_local_next_t;
 
diff --git a/src/vnet/ip/protocols.def b/src/vnet/ip/protocols.def
index 77fab31da05..4ec0bbc9284 100644
--- a/src/vnet/ip/protocols.def
+++ b/src/vnet/ip/protocols.def
@@ -158,5 +158,10 @@ ip_protocol (134, RSVP_E2E_IGNORE)
 ip_protocol (135, MOBILITY)
 ip_protocol (136, UDP_LITE)
 ip_protocol (137, MPLS_IN_IP)
+/*
+ * VPPs way of dealing with fragments is to mark them as 0xfe, to be picked
+ * by the corresponding nodes based on this protocol
+ */
+ip_protocol (0xfe, VPP_FRAGMENTATION)
 ip_protocol (255, RESERVED)