Evolving SRv6 (Segment Routing for IPv6)

Implements:
1.- SR Policies with several (weighted) SID lists
2.- Binding SID
3.- SR LocalSIDs with support for the following functions
	- End
	- End.X
	- End.DX6
	- End.DX4
	- End.DX2
	- End.DT6
	- End.DT2
	- End.B6
	- End.B6.Encaps
4.- SR Steering policies (to steer a traffic through an SR Policy)
	- Support for IPv6 traffic (IPv6 Encapsulation / SRH insertion)
	- Support for IPv4 traffic (IPv6 Encapsulation)
	- Support for L2 traffic
	(T.Insert / T.Encaps)
5.- Doxygen documentation
6.- Framework (APIs) to allow the definition of new SR LocalSID behaviors by means of plugins
7.- Sample SRv6 LocalSID plugin

Change-Id: I2de3d126699d4f11f54c0f7f3b71420ea41fd389
Signed-off-by: Pablo Camarillo <pcamaril@cisco.com>

1 files changed, 568 insertions, 0 deletions

diff --git a/src/vnet/sr/sr_steering.c b/src/vnet/sr/sr_steering.c
new file mode 100755
index 00000000000..86d6f27cbe1
--- /dev/null
+++ b/src/vnet/sr/sr_steering.c
@@ -0,0 +1,568 @@
+/*
+ * sr_steering.c: ipv6 segment routing steering into SR policy
+ *
+ * Copyright (c) 2016 Cisco and/or its affiliates.
+ * Licensed under the Apache License, Version 2.0 (the "License");
+ * you may not use this file except in compliance with the License.
+ * You may obtain a copy of the License at:
+ *
+ *     http://www.apache.org/licenses/LICENSE-2.0
+ *
+ * Unless required by applicable law or agreed to in writing, software
+ * distributed under the License is distributed on an "AS IS" BASIS,
+ * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+ * See the License for the specific language governing permissions and
+ * limitations under the License.
+ */
+
+/**
+ * @file
+ * @brief Packet steering into SR Policies
+ *
+ * This file is in charge of handling the FIB appropiatly to steer packets
+ * through SR Policies as defined in 'sr_policy_rewrite.c'. Notice that here
+ * we are only doing steering. SR policy application is done in
+ * sr_policy_rewrite.c
+ *
+ * Supports:
+ *  - Steering of IPv6 traffic Destination Address based
+ *  - Steering of IPv4 traffic Destination Address based
+ *  - Steering of L2 frames, interface based (sw interface)
+ */
+
+#include <vlib/vlib.h>
+#include <vnet/vnet.h>
+#include <vnet/sr/sr.h>
+#include <vnet/ip/ip.h>
+#include <vnet/sr/sr_packet.h>
+#include <vnet/ip/ip6_packet.h>
+#include <vnet/fib/ip6_fib.h>
+#include <vnet/dpo/dpo.h>
+
+#include <vppinfra/error.h>
+#include <vppinfra/elog.h>
+
+/**
+ * @brief Steer traffic L2 and L3 traffic through a given SR policy
+ *
+ * @param is_del
+ * @param bsid is the bindingSID of the SR Policy (alt to sr_policy_index)
+ * @param sr_policy is the index of the SR Policy (alt to bsid)
+ * @param table_id is the VRF where to install the FIB entry for the BSID
+ * @param prefix is the IPv4/v6 address for L3 traffic type
+ * @param mask_width is the mask for L3 traffic type
+ * @param sw_if_index is the incoming interface for L2 traffic
+ * @param traffic_type describes the type of traffic
+ *
+ * @return 0 if correct, else error
+ */
+int
+sr_steering_policy (int is_del, ip6_address_t * bsid, u32 sr_policy_index,
+		    u32 table_id, ip46_address_t * prefix, u32 mask_width,
+		    u32 sw_if_index, u8 traffic_type)
+{
+  ip6_sr_main_t *sm = &sr_main;
+  sr_steering_key_t key, *key_copy;
+  ip6_sr_steering_policy_t *steer_pl;
+  fib_prefix_t pfx = { 0 };
+
+  ip6_sr_policy_t *sr_policy = 0;
+  uword *p = 0;
+
+  hash_pair_t *hp;
+
+  /* Compute the steer policy key */
+  if (prefix)
+    {
+      key.l3.prefix.as_u64[0] = prefix->as_u64[0];
+      key.l3.prefix.as_u64[1] = prefix->as_u64[1];
+      key.l3.mask_width = mask_width;
+      key.l3.fib_table = (table_id != (u32) ~ 0 ? table_id : 0);
+    }
+  else
+    {
+      key.l2.sw_if_index = sw_if_index;
+
+      /* Sanitise the SW_IF_INDEX */
+      if (pool_is_free_index (sm->vnet_main->interface_main.sw_interfaces,
+			      sw_if_index))
+	return -3;
+
+      vnet_sw_interface_t *sw =
+	vnet_get_sw_interface (sm->vnet_main, sw_if_index);
+      if (sw->type != VNET_SW_INTERFACE_TYPE_HARDWARE)
+	return -3;
+    }
+
+  key.traffic_type = traffic_type;
+
+  /* Search for the item */
+  p = hash_get_mem (sm->steer_policies_index_by_key, &key);
+
+  if (p)
+    {
+      /* Retrieve Steer Policy function */
+      steer_pl = pool_elt_at_index (sm->steer_policies, p[0]);
+
+      if (is_del)
+	{
+	  if (steer_pl->classify.traffic_type == SR_STEER_IPV6)
+	    {
+	      /* Remove FIB entry */
+	      pfx.fp_proto = FIB_PROTOCOL_IP6;
+	      pfx.fp_len = steer_pl->classify.l3.mask_width;
+	      pfx.fp_addr.ip6 = steer_pl->classify.l3.prefix.ip6;
+
+	      fib_table_entry_delete (fib_table_id_find_fib_index
+				      (FIB_PROTOCOL_IP6,
+				       steer_pl->classify.l3.fib_table), &pfx,
+				      FIB_SOURCE_SR);
+	    }
+	  else if (steer_pl->classify.traffic_type == SR_STEER_IPV4)
+	    {
+	      /* Remove FIB entry */
+	      pfx.fp_proto = FIB_PROTOCOL_IP4;
+	      pfx.fp_len = steer_pl->classify.l3.mask_width;
+	      pfx.fp_addr.ip4 = steer_pl->classify.l3.prefix.ip4;
+
+	      fib_table_entry_delete (fib_table_id_find_fib_index
+				      (FIB_PROTOCOL_IP4,
+				       steer_pl->classify.l3.fib_table), &pfx,
+				      FIB_SOURCE_SR);
+	    }
+	  else if (steer_pl->classify.traffic_type == SR_STEER_L2)
+	    {
+	      /* Remove HW redirection */
+	      vnet_feature_enable_disable ("device-input",
+					   "sr-policy-rewrite-encaps-l2",
+					   sw_if_index, 0, 0, 0);
+	      sm->sw_iface_sr_policies[sw_if_index] = ~(u32) 0;
+
+	      /* Remove promiscous mode from interface */
+	      vnet_main_t *vnm = vnet_get_main ();
+	      ethernet_main_t *em = &ethernet_main;
+	      ethernet_interface_t *eif =
+		ethernet_get_interface (em, sw_if_index);
+
+	      if (!eif)
+		goto cleanup_error_redirection;
+
+	      ethernet_set_flags (vnm, sw_if_index, 0);
+	    }
+
+	  /* Delete SR steering policy entry */
+	  pool_put (sm->steer_policies, steer_pl);
+	  hp = hash_get_pair (sm->steer_policies_index_by_key, &key);
+	  key_copy = (void *) (hp->key);
+	  hash_unset_mem (sm->steer_policies_index_by_key, &key);
+	  vec_free (key_copy);
+	  return 1;
+	}
+      else			/* It means user requested to update an existing SR steering policy */
+	{
+	  /* Retrieve SR steering policy */
+	  if (bsid)
+	    {
+	      p = hash_get_mem (sm->sr_policy_index_by_key, bsid);
+	      if (p)
+		sr_policy = pool_elt_at_index (sm->sr_policies, p[0]);
+	      else
+		return -2;
+	    }
+	  else
+	    sr_policy = pool_elt_at_index (sm->sr_policies, sr_policy_index);
+
+	  if (!sr_policy)
+	    return -2;
+
+	  steer_pl->sr_policy = sr_policy - sm->sr_policies;
+
+	  /* Remove old FIB/hw redirection and create a new one */
+	  if (steer_pl->classify.traffic_type == SR_STEER_IPV6)
+	    {
+	      /* Remove FIB entry */
+	      pfx.fp_proto = FIB_PROTOCOL_IP6;
+	      pfx.fp_len = steer_pl->classify.l3.mask_width;
+	      pfx.fp_addr.ip6 = steer_pl->classify.l3.prefix.ip6;
+
+	      fib_table_entry_delete (fib_table_id_find_fib_index
+				      (FIB_PROTOCOL_IP6,
+				       steer_pl->classify.l3.fib_table), &pfx,
+				      FIB_SOURCE_SR);
+
+	      /* Create a new one */
+	      goto update_fib;
+
+	    }
+	  else if (steer_pl->classify.traffic_type == SR_STEER_IPV4)
+	    {
+	      /* Remove FIB entry */
+	      pfx.fp_proto = FIB_PROTOCOL_IP4;
+	      pfx.fp_len = steer_pl->classify.l3.mask_width;
+	      pfx.fp_addr.ip4 = steer_pl->classify.l3.prefix.ip4;
+
+	      fib_table_entry_delete (fib_table_id_find_fib_index
+				      (FIB_PROTOCOL_IP4,
+				       steer_pl->classify.l3.fib_table), &pfx,
+				      FIB_SOURCE_SR);
+
+	      /* Create a new one */
+	      goto update_fib;
+	    }
+	  else if (steer_pl->classify.traffic_type == SR_STEER_L2)
+	    {
+	      /* Update L2-HW redirection */
+	      goto update_fib;
+	    }
+	}
+    }
+  else
+    /* delete; steering policy does not exist; complain */
+  if (is_del)
+    return -4;
+
+  /* Retrieve SR policy */
+  if (bsid)
+    {
+      p = hash_get_mem (sm->sr_policy_index_by_key, bsid);
+      if (p)
+	sr_policy = pool_elt_at_index (sm->sr_policies, p[0]);
+      else
+	return -2;
+    }
+  else
+    sr_policy = pool_elt_at_index (sm->sr_policies, sr_policy_index);
+
+  /* Create a new steering policy */
+  pool_get (sm->steer_policies, steer_pl);
+  memset (steer_pl, 0, sizeof (*steer_pl));
+
+  if (traffic_type == SR_STEER_IPV4 || traffic_type == SR_STEER_IPV6)
+    {
+      clib_memcpy (&steer_pl->classify.l3.prefix, prefix,
+		   sizeof (ip46_address_t));
+      steer_pl->classify.l3.mask_width = mask_width;
+      steer_pl->classify.l3.fib_table =
+	(table_id != (u32) ~ 0 ? table_id : 0);
+      steer_pl->classify.traffic_type = traffic_type;
+    }
+  else if (traffic_type == SR_STEER_L2)
+    {
+      steer_pl->classify.l2.sw_if_index = sw_if_index;
+      steer_pl->classify.traffic_type = traffic_type;
+    }
+  else
+    {
+      /* Incorrect API usage. Should never get here */
+      pool_put (sm->steer_policies, steer_pl);
+      hp = hash_get_pair (sm->steer_policies_index_by_key, &key);
+      key_copy = (void *) (hp->key);
+      hash_unset_mem (sm->steer_policies_index_by_key, &key);
+      vec_free (key_copy);
+      return -1;
+    }
+  steer_pl->sr_policy = sr_policy - sm->sr_policies;
+
+  /* Create and store key */
+  key_copy = vec_new (sr_steering_key_t, 1);
+  clib_memcpy (key_copy, &key, sizeof (sr_steering_key_t));
+  hash_set_mem (sm->steer_policies_index_by_key,
+		key_copy, steer_pl - sm->steer_policies);
+
+  if (traffic_type == SR_STEER_L2)
+    {
+      if (!sr_policy->is_encap)
+	goto cleanup_error_encap;
+
+      if (vnet_feature_enable_disable
+	  ("device-input", "sr-policy-rewrite-encaps-l2", sw_if_index, 1, 0,
+	   0))
+	goto cleanup_error_redirection;
+
+      /* Set promiscous mode on interface */
+      vnet_main_t *vnm = vnet_get_main ();
+      ethernet_main_t *em = &ethernet_main;
+      ethernet_interface_t *eif = ethernet_get_interface (em, sw_if_index);
+
+      if (!eif)
+	goto cleanup_error_redirection;
+
+      ethernet_set_flags (vnm, sw_if_index,
+			  ETHERNET_INTERFACE_FLAG_ACCEPT_ALL);
+    }
+  else if (traffic_type == SR_STEER_IPV4)
+    if (!sr_policy->is_encap)
+      goto cleanup_error_encap;
+
+update_fib:
+  /* FIB API calls - Recursive route through the BindingSID */
+  if (traffic_type == SR_STEER_IPV6)
+    {
+      pfx.fp_proto = FIB_PROTOCOL_IP6;
+      pfx.fp_len = steer_pl->classify.l3.mask_width;
+      pfx.fp_addr.ip6 = steer_pl->classify.l3.prefix.ip6;
+
+      fib_table_entry_path_add (fib_table_id_find_fib_index (FIB_PROTOCOL_IP6,
+							     (table_id !=
+							      (u32) ~ 0 ?
+							      table_id : 0)),
+				&pfx, FIB_SOURCE_CLI, FIB_ENTRY_FLAG_NONE,
+				FIB_PROTOCOL_IP6,
+				(ip46_address_t *) & sr_policy->bsid, ~0,
+				sm->fib_table_ip6, 1, NULL,
+				FIB_ROUTE_PATH_FLAG_NONE);
+    }
+  else if (traffic_type == SR_STEER_IPV4)
+    {
+      pfx.fp_proto = FIB_PROTOCOL_IP4;
+      pfx.fp_len = steer_pl->classify.l3.mask_width;
+      pfx.fp_addr.ip4 = steer_pl->classify.l3.prefix.ip4;
+
+      fib_table_entry_path_add (fib_table_id_find_fib_index (FIB_PROTOCOL_IP4,
+							     (table_id !=
+							      (u32) ~ 0 ?
+							      table_id : 0)),
+				&pfx, FIB_SOURCE_CLI, FIB_ENTRY_FLAG_NONE,
+				FIB_PROTOCOL_IP6,
+				(ip46_address_t *) & sr_policy->bsid, ~0,
+				sm->fib_table_ip4, 1, NULL,
+				FIB_ROUTE_PATH_FLAG_NONE);
+    }
+  else if (traffic_type == SR_STEER_L2)
+    {
+      if (sw_if_index < vec_len (sm->sw_iface_sr_policies))
+	sm->sw_iface_sr_policies[sw_if_index] = steer_pl->sr_policy;
+      else
+	{
+	  vec_resize (sm->sw_iface_sr_policies,
+		      (pool_len (sm->vnet_main->interface_main.sw_interfaces)
+		       - vec_len (sm->sw_iface_sr_policies)));
+	  sm->sw_iface_sr_policies[sw_if_index] = steer_pl->sr_policy;
+	}
+    }
+
+  return 0;
+
+cleanup_error_encap:
+  pool_put (sm->steer_policies, steer_pl);
+  hp = hash_get_pair (sm->steer_policies_index_by_key, &key);
+  key_copy = (void *) (hp->key);
+  hash_unset_mem (sm->steer_policies_index_by_key, &key);
+  vec_free (key_copy);
+  return -5;
+
+cleanup_error_redirection:
+  pool_put (sm->steer_policies, steer_pl);
+  hp = hash_get_pair (sm->steer_policies_index_by_key, &key);
+  key_copy = (void *) (hp->key);
+  hash_unset_mem (sm->steer_policies_index_by_key, &key);
+  vec_free (key_copy);
+  return -3;
+}
+
+static clib_error_t *
+sr_steer_policy_command_fn (vlib_main_t * vm, unformat_input_t * input,
+			    vlib_cli_command_t * cmd)
+{
+  vnet_main_t *vnm = vnet_get_main ();
+
+  int is_del = 0;
+
+  ip46_address_t prefix;
+  u32 dst_mask_width = 0;
+  u32 sw_if_index = (u32) ~ 0;
+  u8 traffic_type = 0;
+  u32 fib_table = (u32) ~ 0;
+
+  ip6_address_t bsid;
+  u32 sr_policy_index = (u32) ~ 0;
+
+  u8 sr_policy_set = 0;
+
+  int rv;
+  while (unformat_check_input (input) != UNFORMAT_END_OF_INPUT)
+    {
+      if (unformat (input, "del"))
+	is_del = 1;
+      else if (!traffic_type
+	       && unformat (input, "l3 %U/%d", unformat_ip6_address,
+			    &prefix.ip6, &dst_mask_width))
+	traffic_type = SR_STEER_IPV6;
+      else if (!traffic_type
+	       && unformat (input, "l3 %U/%d", unformat_ip4_address,
+			    &prefix.ip4, &dst_mask_width))
+	traffic_type = SR_STEER_IPV4;
+      else if (!traffic_type
+	       && unformat (input, "l2 %U", unformat_vnet_sw_interface, vnm,
+			    &sw_if_index))
+	traffic_type = SR_STEER_L2;
+      else if (!sr_policy_set
+	       && unformat (input, "via sr policy index %d",
+			    &sr_policy_index))
+	sr_policy_set = 1;
+      else if (!sr_policy_set
+	       && unformat (input, "via sr policy bsid %U",
+			    unformat_ip6_address, &bsid))
+	sr_policy_set = 1;
+      else if (fib_table == (u32) ~ 0
+	       && unformat (input, "fib-table %d", &fib_table));
+      else
+	break;
+    }
+
+  if (!traffic_type)
+    return clib_error_return (0, "No L2/L3 traffic specified");
+  if (!sr_policy_set)
+    return clib_error_return (0, "No SR policy specified");
+
+  /* Make sure that the prefixes are clean */
+  if (traffic_type == SR_STEER_IPV4)
+    {
+      u32 mask =
+	(dst_mask_width ? (0xFFFFFFFFu >> (32 - dst_mask_width)) : 0);
+      prefix.ip4.as_u32 &= mask;
+    }
+  else if (traffic_type == SR_STEER_IPV6)
+    {
+      ip6_address_t mask;
+      ip6_address_mask_from_width (&mask, dst_mask_width);
+      ip6_address_mask (&prefix.ip6, &mask);
+    }
+
+  rv =
+    sr_steering_policy (is_del, (sr_policy_index == ~(u32) 0 ? &bsid : NULL),
+			sr_policy_index, fib_table, &prefix, dst_mask_width,
+			sw_if_index, traffic_type);
+
+  switch (rv)
+    {
+    case 0:
+      break;
+    case 1:
+      return 0;
+    case -1:
+      return clib_error_return (0, "Incorrect API usage.");
+    case -2:
+      return clib_error_return (0,
+				"The requested SR policy could not be located. Review the BSID/index.");
+    case -3:
+      return clib_error_return (0,
+				"Unable to do SW redirect. Incorrect interface.");
+    case -4:
+      return clib_error_return (0,
+				"The requested SR policy could not be deleted. Review the BSID/index.");
+    case -5:
+      return clib_error_return (0,
+				"The SR policy is not an encapsulation one.");
+    default:
+      return clib_error_return (0, "BUG: sr steer policy returns %d", rv);
+    }
+  return 0;
+}
+
+/* *INDENT-OFF* */
+VLIB_CLI_COMMAND (sr_steer_policy_command, static) = {
+  .path = "sr steer",
+  .short_help = "sr steer (del) [l3 <ip_addr/mask>|l2 <sf_if>]"
+    "via sr policy [index <sr_policy_index>|bsid <bsid_ip6_addr>]"
+    "(fib-table <fib_table_index>)",
+  .long_help =
+    "\tSteer a L2 or L3 traffic through an existing SR policy.\n"
+    "\tExamples:\n"
+    "\t\tsr steer l3 2001::/64 via sr_policy index 5\n"
+    "\t\tsr steer l3 2001::/64 via sr_policy bsid 2010::9999:1\n"
+    "\t\tsr steer l2 GigabitEthernet0/5/0 via sr_policy index 5\n"
+    "\t\tsr steer del l3 2001::/64 via sr_policy index 5\n",
+  .function = sr_steer_policy_command_fn,
+};
+/* *INDENT-ON* */
+
+static clib_error_t *
+show_sr_steering_policies_command_fn (vlib_main_t * vm,
+				      unformat_input_t * input,
+				      vlib_cli_command_t * cmd)
+{
+  ip6_sr_main_t *sm = &sr_main;
+  ip6_sr_steering_policy_t **steer_policies = 0;
+  ip6_sr_steering_policy_t *steer_pl;
+
+  vnet_main_t *vnm = vnet_get_main ();
+
+  ip6_sr_policy_t *pl = 0;
+  int i;
+
+  vlib_cli_output (vm, "SR steering policies:");
+  /* *INDENT-OFF* */
+  pool_foreach (steer_pl, sm->steer_policies, ({vec_add1(steer_policies, steer_pl);}));
+  /* *INDENT-ON* */
+  vlib_cli_output (vm, "Traffic\t\tSR policy BSID");
+  for (i = 0; i < vec_len (steer_policies); i++)
+    {
+      steer_pl = steer_policies[i];
+      pl = pool_elt_at_index (sm->sr_policies, steer_pl->sr_policy);
+      if (steer_pl->classify.traffic_type == SR_STEER_L2)
+	{
+	  vlib_cli_output (vm, "L2 %U\t%U",
+			   format_vnet_sw_if_index_name, vnm,
+			   steer_pl->classify.l2.sw_if_index,
+			   format_ip6_address, &pl->bsid);
+	}
+      else if (steer_pl->classify.traffic_type == SR_STEER_IPV4)
+	{
+	  vlib_cli_output (vm, "L3 %U/%d\t%U",
+			   format_ip4_address,
+			   &steer_pl->classify.l3.prefix.ip4,
+			   steer_pl->classify.l3.mask_width,
+			   format_ip6_address, &pl->bsid);
+	}
+      else if (steer_pl->classify.traffic_type == SR_STEER_IPV6)
+	{
+	  vlib_cli_output (vm, "L3 %U/%d\t%U",
+			   format_ip6_address,
+			   &steer_pl->classify.l3.prefix.ip6,
+			   steer_pl->classify.l3.mask_width,
+			   format_ip6_address, &pl->bsid);
+	}
+    }
+  return 0;
+}
+
+/* *INDENT-OFF* */
+VLIB_CLI_COMMAND (show_sr_steering_policies_command, static) = {
+  .path = "show sr steering policies",
+  .short_help = "show sr steering policies",
+  .function = show_sr_steering_policies_command_fn,
+};
+/* *INDENT-ON* */
+
+clib_error_t *
+sr_steering_init (vlib_main_t * vm)
+{
+  ip6_sr_main_t *sm = &sr_main;
+
+  /* Init memory for function keys */
+  sm->steer_policies_index_by_key =
+    hash_create_mem (0, sizeof (sr_steering_key_t), sizeof (uword));
+
+  sm->sw_iface_sr_policies = 0;
+
+  sm->vnet_main = vnet_get_main ();
+
+  return 0;
+}
+
+VLIB_INIT_FUNCTION (sr_steering_init);
+
+VNET_FEATURE_INIT (sr_policy_rewrite_encaps_l2, static) =
+{
+.arc_name = "device-input",.node_name =
+    "sr-pl-rewrite-encaps-l2",.runs_before =
+    VNET_FEATURES ("ethernet-input"),};
+
+/*
+* fd.io coding-style-patch-verification: ON
+*
+* Local Variables:
+* eval: (c-set-style "gnu")
+* End:
+*/