/src/frr/ospfd/ospf_sr.c

Source
// SPDX-License-Identifier: GPL-2.0-or-later
/*
 * This is an implementation of Segment Routing
 * as per RFC 8665 - OSPF Extensions for Segment Routing
 * and RFC 8476 - Signaling Maximum SID Depth (MSD) Using OSPF
 *
 * Module name: Segment Routing
 *
 * Author: Olivier Dugeon <olivier.dugeon@orange.com>
 * Author: Anselme Sawadogo <anselmesawadogo@gmail.com>
 *
 * Copyright (C) 2016 - 2020 Orange Labs http://www.orange.com
 */

#ifdef HAVE_CONFIG_H
#include "config.h"
#endif

#include <math.h>
#include <stdio.h>
#include <stdlib.h>
#include <zebra.h>

#include "printfrr.h"
#include "command.h"
#include "hash.h"
#include "if.h"
#include "if.h"
#include "jhash.h"
#include "libospf.h" /* for ospf interface types */
#include "linklist.h"
#include "log.h"
#include "memory.h"
#include "monotime.h"
#include "network.h"
#include "prefix.h"
#include "sockunion.h" /* for inet_aton() */
#include "stream.h"
#include "table.h"
#include "frrevent.h"
#include "vty.h"
#include "zclient.h"
#include "sbuf.h"
#include <lib/json.h>
#include "ospf_errors.h"

#include "ospfd/ospfd.h"
#include "ospfd/ospf_interface.h"
#include "ospfd/ospf_ism.h"
#include "ospfd/ospf_asbr.h"
#include "ospfd/ospf_lsa.h"
#include "ospfd/ospf_lsdb.h"
#include "ospfd/ospf_neighbor.h"
#include "ospfd/ospf_nsm.h"
#include "ospfd/ospf_flood.h"
#include "ospfd/ospf_packet.h"
#include "ospfd/ospf_spf.h"
#include "ospfd/ospf_dump.h"
#include "ospfd/ospf_route.h"
#include "ospfd/ospf_ase.h"
#include "ospfd/ospf_sr.h"
#include "ospfd/ospf_ri.h"
#include "ospfd/ospf_ext.h"
#include "ospfd/ospf_zebra.h"

/*
 * Global variable to manage Segment Routing on this node.
 * Note that all parameter values are stored in network byte order.
 */
static struct ospf_sr_db OspfSR;
static void ospf_sr_register_vty(void);
static inline void del_adj_sid(struct sr_nhlfe nhlfe);
static int ospf_sr_start(struct ospf *ospf);

/*
 * Segment Routing Data Base functions
 */

/* Hash function for Segment Routing entry */
static unsigned int sr_hash(const void *p)
{
  const struct in_addr *rid = p;

  return jhash_1word(rid->s_addr, 0);
}

/* Compare 2 Router ID hash entries based on SR Node */
static bool sr_cmp(const void *p1, const void *p2)
{
  const struct sr_node *srn = p1;
  const struct in_addr *rid = p2;

  return IPV4_ADDR_SAME(&srn->adv_router, rid);
}

/* Functions to remove an SR Link */
static void del_sr_link(void *val)
{
  struct sr_link *srl = (struct sr_link *)val;

  del_adj_sid(srl->nhlfe[0]);
  del_adj_sid(srl->nhlfe[1]);
  XFREE(MTYPE_OSPF_SR_PARAMS, val);
}

/* Functions to remove an SR Prefix */
static void del_sr_pref(void *val)
{
  struct sr_prefix *srp = (struct sr_prefix *)val;

  ospf_zebra_delete_prefix_sid(srp);
  XFREE(MTYPE_OSPF_SR_PARAMS, val);
}

/* Allocate new Segment Routine node */
static struct sr_node *sr_node_new(struct in_addr *rid)
{

  if (rid == NULL)
    return NULL;

  struct sr_node *new;

  /* Allocate Segment Routing node memory */
  new = XCALLOC(MTYPE_OSPF_SR_PARAMS, sizeof(struct sr_node));

  /* Default Algorithm, SRGB and MSD */
  for (int i = 0; i < ALGORITHM_COUNT; i++)
    new->algo[i] = SR_ALGORITHM_UNSET;

  new->srgb.range_size = 0;
  new->srgb.lower_bound = 0;
  new->msd = 0;

  /* Create Link, Prefix and Range TLVs list */
  new->ext_link = list_new();
  new->ext_prefix = list_new();
  new->ext_link->del = del_sr_link;
  new->ext_prefix->del = del_sr_pref;

  IPV4_ADDR_COPY(&new->adv_router, rid);
  new->neighbor = NULL;
  new->instance = 0;

  osr_debug("  |-  Created new SR node for %pI4", &new->adv_router);
  return new;
}

/* Supposed to be used for testing */
struct sr_node *ospf_sr_node_create(struct in_addr *rid)
{
  struct sr_node *srn;

  srn = hash_get(OspfSR.neighbors, (void *)rid, (void *)sr_node_new);

  return srn;
}

/* Delete Segment Routing node */
static void sr_node_del(struct sr_node *srn)
{
  /* Sanity Check */
  if (srn == NULL)
    return;

  osr_debug("  |- Delete SR node for %pI4", &srn->adv_router);

  /* Clean Extended Link */
  list_delete(&srn->ext_link);

  /* Clean Prefix List */
  list_delete(&srn->ext_prefix);

  XFREE(MTYPE_OSPF_SR_PARAMS, srn);
}

/* Get SR Node for a given nexthop */
static struct sr_node *get_sr_node_by_nexthop(struct ospf *ospf,
                struct in_addr nexthop)
{
  struct ospf_interface *oi = NULL;
  struct ospf_neighbor *nbr = NULL;
  struct listnode *node;
  struct route_node *rn;
  struct sr_node *srn;
  bool found;

  /* Sanity check */
  if (OspfSR.neighbors == NULL)
    return NULL;

  osr_debug("      |-  Search SR-Node for nexthop %pI4", &nexthop);

  /* First, search neighbor Router ID for this nexthop */
  found = false;
  for (ALL_LIST_ELEMENTS_RO(ospf->oiflist, node, oi)) {
    for (rn = route_top(oi->nbrs); rn; rn = route_next(rn)) {
      nbr = rn->info;
      if ((nbr) && (IPV4_ADDR_SAME(&nexthop, &nbr->src))) {
        found = true;
        break;
      }
    }
    if (found)
      break;
  }

  if (!found)
    return NULL;

  osr_debug("      |-  Found nexthop Router ID %pI4", &nbr->router_id);

  /* Then, search SR Node */
  srn = (struct sr_node *)hash_lookup(OspfSR.neighbors, &nbr->router_id);

  return srn;
}

/*
 * Segment Routing Local Block management functions
 */

/**
 * It is necessary to known which label is already allocated to manage the range
 * of SRLB. This is particular useful when an interface flap (goes up / down
 * frequently). Here, SR will release and then allocate label for the Adjacency
 * for each concerned interface. If we don't care, there is a risk to run out of
 * label.
 *
 * For that purpose, a similar principle as already provided to manage chunk of
 * label is proposed. But, here, the label chunk has not a fix range of 64
 * labels that could be easily manage with a single variable of 64 bits size.
 * So, used_mark is used as a bit wise to mark label reserved (bit set) or not
 * (bit unset). Its size is equal to the number of label of the SRLB range round
 * up to 64 bits.
 *
 *  - sr__local_block_init() computes the number of 64 bits variables that are
 *    needed to manage the SRLB range and allocates this number.
 *  - ospf_sr_local_block_request_label() pick up the first available label and
 *    set corresponding bit
 *  - ospf_sr_local_block_release_label() release label by reseting the
 *    corresponding bit and set the next label to the first free position
 */

/**
 * Initialize Segment Routing Local Block from SRDB configuration and reserve
 * block of bits to manage label allocation.
 *
 * @param lower_bound The lower bound of the SRLB range
 * @param upper_bound The upper bound of the SRLB range
 *
 * @return    0 on success, -1 otherwise
 */
static int sr_local_block_init(uint32_t lower_bound, uint32_t upper_bound)
{
  struct sr_local_block *srlb = &OspfSR.srlb;
  uint32_t size;

  /* Check if SRLB is not already configured */
  if (srlb->reserved)
    return 0;

  /*
   * Request SRLB to the label manager. If the allocation fails, return
   * an error to disable SR until a new SRLB is successfully allocated.
   */
  size = upper_bound - lower_bound + 1;
  if (ospf_zebra_request_label_range(lower_bound, size)) {
    zlog_err("SR: Error reserving SRLB [%u/%u] %u labels",
       lower_bound, upper_bound, size);
    return -1;
  }

  osr_debug("SR: Got new SRLB [%u/%u], %u labels", lower_bound,
      upper_bound, size);

  /* Initialize the SRLB */
  srlb->start = lower_bound;
  srlb->end = upper_bound;
  srlb->current = 0;

  /* Compute the needed Used Mark number and allocate them */
  srlb->max_block = size / SRLB_BLOCK_SIZE;
  if ((size % SRLB_BLOCK_SIZE) != 0)
    srlb->max_block++;
  srlb->used_mark = XCALLOC(MTYPE_OSPF_SR_PARAMS,
          srlb->max_block * SRLB_BLOCK_SIZE);
  srlb->reserved = true;

  return 0;
}

static int sr_global_block_init(uint32_t start, uint32_t size)
{
  struct sr_global_block *srgb = &OspfSR.srgb;

  /* Check if already configured */
  if (srgb->reserved)
    return 0;

  /* request chunk */
  uint32_t end = start + size - 1;
  if (ospf_zebra_request_label_range(start, size) < 0) {
    zlog_err("SR: Error reserving SRGB [%u/%u], %u labels", start,
       end, size);
    return -1;
  }

  osr_debug("SR: Got new SRGB [%u/%u], %u labels", start, end, size);

  /* success */
  srgb->start = start;
  srgb->size = size;
  srgb->reserved = true;
  return 0;
}

/**
 * Remove Segment Routing Local Block.
 *
 */
static void sr_local_block_delete(void)
{
  struct sr_local_block *srlb = &OspfSR.srlb;

  /* Check if SRLB is not already delete */
  if (!srlb->reserved)
    return;

  osr_debug("SR (%s): Remove SRLB [%u/%u]", __func__, srlb->start,
      srlb->end);

  /* First release the label block */
  ospf_zebra_release_label_range(srlb->start, srlb->end);

  /* Then reset SRLB structure */
  if (srlb->used_mark != NULL)
    XFREE(MTYPE_OSPF_SR_PARAMS, srlb->used_mark);

  srlb->reserved = false;
}

/**
 * Remove Segment Routing Global block
 */
static void sr_global_block_delete(void)
{
  struct sr_global_block *srgb = &OspfSR.srgb;

  if (!srgb->reserved)
    return;

  osr_debug("SR (%s): Remove SRGB [%u/%u]", __func__, srgb->start,
      srgb->start + srgb->size - 1);

  ospf_zebra_release_label_range(srgb->start,
               srgb->start + srgb->size - 1);

  srgb->reserved = false;
}


/**
 * Request a label from the Segment Routing Local Block.
 *
 * @return  First available label on success or MPLS_INVALID_LABEL if the
 *    block of labels is full
 */
mpls_label_t ospf_sr_local_block_request_label(void)
{
  struct sr_local_block *srlb = &OspfSR.srlb;
  mpls_label_t label;
  uint32_t index;
  uint32_t pos;
  uint32_t size = srlb->end - srlb->start + 1;

  /* Check if we ran out of available labels */
  if (srlb->current >= size)
    return MPLS_INVALID_LABEL;

  /* Get first available label and mark it used */
  label = srlb->current + srlb->start;
  index = srlb->current / SRLB_BLOCK_SIZE;
  pos = 1ULL << (srlb->current % SRLB_BLOCK_SIZE);
  srlb->used_mark[index] |= pos;

  /* Jump to the next free position */
  srlb->current++;
  pos = srlb->current % SRLB_BLOCK_SIZE;
  while (srlb->current < size) {
    if (pos == 0)
      index++;
    if (!((1ULL << pos) & srlb->used_mark[index]))
      break;
    else {
      srlb->current++;
      pos = srlb->current % SRLB_BLOCK_SIZE;
    }
  }

  if (srlb->current == size)
    zlog_warn(
      "SR: Warning, SRLB is depleted and next label request will fail");

  return label;
}

/**
 * Release label in the Segment Routing Local Block.
 *
 * @param label Label to be release
 *
 * @return  0 on success or -1 if label falls outside SRLB
 */
int ospf_sr_local_block_release_label(mpls_label_t label)
{
  struct sr_local_block *srlb = &OspfSR.srlb;
  uint32_t index;
  uint32_t pos;

  /* Check that label falls inside the SRLB */
  if ((label < srlb->start) || (label > srlb->end)) {
    flog_warn(EC_OSPF_SR_SID_OVERFLOW,
        "%s: Returning label %u is outside SRLB [%u/%u]",
        __func__, label, srlb->start, srlb->end);
    return -1;
  }

  index = (label - srlb->start) / SRLB_BLOCK_SIZE;
  pos = 1ULL << ((label - srlb->start) % SRLB_BLOCK_SIZE);
  srlb->used_mark[index] &= ~pos;
  /* Reset current to the first available position */
  for (index = 0; index < srlb->max_block; index++) {
    if (srlb->used_mark[index] != 0xFFFFFFFFFFFFFFFF) {
      for (pos = 0; pos < SRLB_BLOCK_SIZE; pos++)
        if (!((1ULL << pos) & srlb->used_mark[index])) {
          srlb->current =
            index * SRLB_BLOCK_SIZE + pos;
          break;
        }
      break;
    }
  }

  return 0;
}

/*
 * Segment Routing Initialization functions
 */

/**
 * Thread function to re-attempt connection to the Label Manager and thus be
 * able to start Segment Routing.
 *
 * @param start   Thread structure that contains area as argument
 *
 * @return    1 on success
 */
static void sr_start_label_manager(struct event *start)
{
  struct ospf *ospf;

  ospf = EVENT_ARG(start);

  /* re-attempt to start SR & Label Manager connection */
  ospf_sr_start(ospf);
}

/* Segment Routing starter function */
static int ospf_sr_start(struct ospf *ospf)
{
  struct route_node *rn;
  struct ospf_lsa *lsa;
  struct sr_node *srn;
  int rc = 0;

  osr_debug("SR (%s): Start Segment Routing", __func__);

  /* Initialize self SR Node if not already done */
  if (OspfSR.self == NULL) {
    srn = hash_get(OspfSR.neighbors, (void *)&(ospf->router_id),
             (void *)sr_node_new);

    /* Complete & Store self SR Node */
    srn->srgb.range_size = OspfSR.srgb.size;
    srn->srgb.lower_bound = OspfSR.srgb.start;
    srn->srlb.lower_bound = OspfSR.srlb.start;
    srn->srlb.range_size = OspfSR.srlb.end - OspfSR.srlb.start + 1;
    srn->algo[0] = OspfSR.algo[0];
    srn->msd = OspfSR.msd;
    OspfSR.self = srn;
  }

  /* Then, start Label Manager if not ready */
  if (!ospf_zebra_label_manager_ready())
    if (ospf_zebra_label_manager_connect() < 0) {
      /* Re-attempt to connect to Label Manager in 1 sec. */
      event_add_timer(master, sr_start_label_manager, ospf, 1,
          &OspfSR.t_start_lm);
      osr_debug("  |- Failed to start the Label Manager");
      return -1;
    }

  /*
   * Request SRLB & SGRB to the label manager if not already reserved.
   * If the allocation fails, return an error to disable SR until a new
   * SRLB and/or SRGB are successfully allocated.
   */
  if (sr_local_block_init(OspfSR.srlb.start, OspfSR.srlb.end) < 0)
    return -1;

  if (sr_global_block_init(OspfSR.srgb.start, OspfSR.srgb.size) < 0)
    return -1;

  /* SR is UP and ready to flood LSA */
  OspfSR.status = SR_UP;

  /* Set Router Information SR parameters */
  osr_debug("SR: Activate SR for Router Information LSA");

  ospf_router_info_update_sr(true, OspfSR.self);

  /* Update Ext LSA */
  osr_debug("SR: Activate SR for Extended Link/Prefix LSA");

  ospf_ext_update_sr(true);

  osr_debug("SR (%s): Update SR-DB from LSDB", __func__);

  /* Start by looking to Router Info & Extended LSA in lsdb */
  if ((ospf != NULL) && (ospf->backbone != NULL)) {
    LSDB_LOOP (OPAQUE_AREA_LSDB(ospf->backbone), rn, lsa) {
      if (IS_LSA_MAXAGE(lsa) || IS_LSA_SELF(lsa))
        continue;
      int lsa_id =
        GET_OPAQUE_TYPE(ntohl(lsa->data->id.s_addr));
      switch (lsa_id) {
      case OPAQUE_TYPE_ROUTER_INFORMATION_LSA:
        ospf_sr_ri_lsa_update(lsa);
        break;
      case OPAQUE_TYPE_EXTENDED_PREFIX_LSA:
        ospf_sr_ext_prefix_lsa_update(lsa);
        break;
      case OPAQUE_TYPE_EXTENDED_LINK_LSA:
        ospf_sr_ext_link_lsa_update(lsa);
        break;
      default:
        break;
      }
    }
  }

  rc = 1;
  return rc;
}

/* Stop Segment Routing */
static void ospf_sr_stop(void)
{

  if (OspfSR.status == SR_OFF)
    return;

  osr_debug("SR (%s): Stop Segment Routing", __func__);

  /* Disable any re-attempt to connect to Label Manager */
  EVENT_OFF(OspfSR.t_start_lm);

  /* Release SRGB if active */
  sr_global_block_delete();

  /* Release SRLB if active */
  sr_local_block_delete();

  /*
   * Remove all SR Nodes from the Hash table. Prefix and Link SID will
   * be remove though list_delete() call. See sr_node_del()
   */
  hash_clean(OspfSR.neighbors, (void *)sr_node_del);
  OspfSR.self = NULL;
  OspfSR.status = SR_OFF;
}

/*
 * Segment Routing initialize function
 *
 * @param - nothing
 *
 * @return 0 if OK, -1 otherwise
 */
int ospf_sr_init(void)
{
  int rc = -1;

  osr_debug("SR (%s): Initialize SR Data Base", __func__);

  memset(&OspfSR, 0, sizeof(OspfSR));
  OspfSR.status = SR_OFF;
  /* Only AREA flooding is supported in this release */
  OspfSR.scope = OSPF_OPAQUE_AREA_LSA;

  /* Initialize Algorithms, SRGB, SRLB and MSD TLVs */
  /* Only Algorithm SPF is supported */
  OspfSR.algo[0] = SR_ALGORITHM_SPF;
  for (int i = 1; i < ALGORITHM_COUNT; i++)
    OspfSR.algo[i] = SR_ALGORITHM_UNSET;

  OspfSR.srgb.size = DEFAULT_SRGB_SIZE;
  OspfSR.srgb.start = DEFAULT_SRGB_LABEL;
  OspfSR.srgb.reserved = false;

  OspfSR.srlb.start = DEFAULT_SRLB_LABEL;
  OspfSR.srlb.end = DEFAULT_SRLB_END;
  OspfSR.srlb.reserved = false;
  OspfSR.msd = 0;

  /* Initialize Hash table for neighbor SR nodes */
  OspfSR.neighbors = hash_create(sr_hash, sr_cmp, "OSPF_SR");
  if (OspfSR.neighbors == NULL)
    return rc;

  /* Register Segment Routing VTY command */
  ospf_sr_register_vty();

  rc = 0;
  return rc;
}

/*
 * Segment Routing termination function
 *
 * @param - nothing
 * @return - nothing
 */
void ospf_sr_term(void)
{

  /* Stop Segment Routing */
  ospf_sr_stop();

  hash_clean_and_free(&OspfSR.neighbors, (void *)sr_node_del);
}

/*
 * Segment Routing finish function
 *
 * @param - nothing
 * @return - nothing
 */
void ospf_sr_finish(void)
{
  /* Stop Segment Routing */
  ospf_sr_stop();
}

/*
 * Following functions are used to manipulate the
 * Next Hop Label Forwarding entry (NHLFE)
 */

/* Compute label from index */
static mpls_label_t index2label(uint32_t index, struct sr_block srgb)
{
  mpls_label_t label;

  label = srgb.lower_bound + index;
  if (label > (srgb.lower_bound + srgb.range_size)) {
    flog_warn(EC_OSPF_SR_SID_OVERFLOW,
        "%s: SID index %u falls outside SRGB range",
        __func__, index);
    return MPLS_INVALID_LABEL;
  } else
    return label;
}

/* Get the prefix sid for a specific router id */
mpls_label_t ospf_sr_get_prefix_sid_by_id(struct in_addr *id)
{
  struct sr_node *srn;
  struct sr_prefix *srp;
  mpls_label_t label;

  srn = (struct sr_node *)hash_lookup(OspfSR.neighbors, id);

  if (srn) {
    /*
     * TODO: Here we assume that the SRGBs are the same,
     * and that the node's prefix SID is at the head of
     * the list, probably needs tweaking.
     */
    srp = listnode_head(srn->ext_prefix);
    label = index2label(srp->sid, srn->srgb);
  } else {
    label = MPLS_INVALID_LABEL;
  }

  return label;
}

/* Get the adjacency sid for a specific 'root' id and 'neighbor' id */
mpls_label_t ospf_sr_get_adj_sid_by_id(struct in_addr *root_id,
               struct in_addr *neighbor_id)
{
  struct sr_node *srn;
  struct sr_link *srl;
  mpls_label_t label;
  struct listnode *node;

  srn = (struct sr_node *)hash_lookup(OspfSR.neighbors, root_id);

  label = MPLS_INVALID_LABEL;

  if (srn) {
    for (ALL_LIST_ELEMENTS_RO(srn->ext_link, node, srl)) {
      if (srl->type == ADJ_SID
          && srl->remote_id.s_addr == neighbor_id->s_addr) {
        label = srl->sid[0];
        break;
      }
    }
  }

  return label;
}

/* Get neighbor full structure from address */
static struct ospf_neighbor *get_neighbor_by_addr(struct ospf *top,
              struct in_addr addr)
{
  struct ospf_neighbor *nbr;
  struct ospf_interface *oi;
  struct listnode *node;
  struct route_node *rn;

  /* Sanity Check */
  if (top == NULL)
    return NULL;

  for (ALL_LIST_ELEMENTS_RO(top->oiflist, node, oi))
    for (rn = route_top(oi->nbrs); rn; rn = route_next(rn)) {
      nbr = rn->info;
      if (!nbr)
        continue;

      if (IPV4_ADDR_SAME(&nbr->address.u.prefix4, &addr) ||
          IPV4_ADDR_SAME(&nbr->router_id, &addr)) {
        route_unlock_node(rn);
        return nbr;
      }
    }
  return NULL;
}

/* Get OSPF Path from address */
static struct ospf_route *get_nexthop_by_addr(struct ospf *top,
                struct prefix_ipv4 p)
{
  struct route_node *rn;

  /* Sanity Check */
  if (top == NULL)
    return NULL;

  osr_debug("      |-  Search Nexthop for prefix %pFX",
      (struct prefix *)&p);

  rn = route_node_lookup(top->new_table, (struct prefix *)&p);

  /*
   * Check if we found an OSPF route. May be NULL if SPF has not
   * yet populate routing table for this prefix.
   */
  if (rn == NULL)
    return NULL;

  route_unlock_node(rn);
  return rn->info;
}

/* Compute NHLFE entry for Extended Link */
static int compute_link_nhlfe(struct sr_link *srl)
{
  struct ospf *top = ospf_lookup_by_vrf_id(VRF_DEFAULT);
  struct ospf_neighbor *nh;
  int rc = 0;

  osr_debug("    |-  Compute NHLFE for link %pI4", &srl->itf_addr);

  /* First determine the OSPF Neighbor */
  nh = get_neighbor_by_addr(top, srl->nhlfe[0].nexthop);

  /* Neighbor could be not found when OSPF Adjacency just fire up
   * because SPF don't yet populate routing table. This NHLFE will
   * be fixed later when SR SPF schedule will be called.
   */
  if (nh == NULL)
    return rc;

  osr_debug("    |-  Found nexthop %pI4", &nh->router_id);

  /* Set ifindex for this neighbor */
  srl->nhlfe[0].ifindex = nh->oi->ifp->ifindex;
  srl->nhlfe[1].ifindex = nh->oi->ifp->ifindex;

  /* Update neighbor address for LAN_ADJ_SID */
  if (srl->type == LAN_ADJ_SID) {
    IPV4_ADDR_COPY(&srl->nhlfe[0].nexthop, &nh->src);
    IPV4_ADDR_COPY(&srl->nhlfe[1].nexthop, &nh->src);
  }

  /* Set Input & Output Label */
  if (CHECK_FLAG(srl->flags[0], EXT_SUBTLV_LINK_ADJ_SID_VFLG))
    srl->nhlfe[0].label_in = srl->sid[0];
  else
    srl->nhlfe[0].label_in =
      index2label(srl->sid[0], srl->srn->srgb);
  if (CHECK_FLAG(srl->flags[1], EXT_SUBTLV_LINK_ADJ_SID_VFLG))
    srl->nhlfe[1].label_in = srl->sid[1];
  else
    srl->nhlfe[1].label_in =
      index2label(srl->sid[1], srl->srn->srgb);

  srl->nhlfe[0].label_out = MPLS_LABEL_IMPLICIT_NULL;
  srl->nhlfe[1].label_out = MPLS_LABEL_IMPLICIT_NULL;

  rc = 1;
  return rc;
}

/**
 * Compute output label for the given Prefix-SID.
 *
 * @param srp   Segment Routing Prefix
 * @param srnext  Segment Routing nexthop node
 *
 * @return    MPLS label or MPLS_INVALID_LABEL in case of error
 */
static mpls_label_t sr_prefix_out_label(const struct sr_prefix *srp,
          const struct sr_node *srnext)
{
  /* Check if the nexthop SR Node is the last hop? */
  if (srnext == srp->srn) {
    /* SR-Node doesn't request NO-PHP. Return Implicit NULL label */
    if (!CHECK_FLAG(srp->flags, EXT_SUBTLV_PREFIX_SID_NPFLG))
      return MPLS_LABEL_IMPLICIT_NULL;

    /* SR-Node requests Explicit NULL Label */
    if (CHECK_FLAG(srp->flags, EXT_SUBTLV_PREFIX_SID_EFLG))
      return MPLS_LABEL_IPV4_EXPLICIT_NULL;
    /* Fallthrough */
  }

  /* Return SID value as MPLS label if it is an Absolute SID */
  if (CHECK_FLAG(srp->flags, EXT_SUBTLV_PREFIX_SID_VFLG
             | EXT_SUBTLV_PREFIX_SID_LFLG)) {
    /*
     * V/L SIDs have local significance, so only adjacent routers
     * can use them (RFC8665 section #5)
     */
    if (srp->srn != srnext)
      return MPLS_INVALID_LABEL;
    return srp->sid;
  }

  /* Return MPLS label as SRGB lower bound + SID index as per RFC 8665 */
  return (index2label(srp->sid, srnext->srgb));
}

/*
 * Compute NHLFE entry for Extended Prefix
 *
 * @param srp - Segment Routing Prefix
 *
 * @return -1 if no route is found, 0 if there is no SR route ready
 *         and 1 if success or update
 */
static int compute_prefix_nhlfe(struct sr_prefix *srp)
{
  struct ospf *top = ospf_lookup_by_vrf_id(VRF_DEFAULT);
  struct ospf_path *path;
  struct listnode *node;
  struct sr_node *srnext;
  int rc = -1;

  osr_debug("    |-  Compute NHLFE for prefix %pFX",
      (struct prefix *)&srp->prefv4);


  /* First determine the nexthop */
  srp->route = get_nexthop_by_addr(top, srp->prefv4);

  /* Nexthop could be not found when OSPF Adjacency just fire up
   * because SPF don't yet populate routing table. This NHLFE will
   * be fixed later when SR SPF schedule will be called.
   */
  if (srp->route == NULL)
    return rc;

  /* Compute Input Label with self SRGB */
  srp->label_in = index2label(srp->sid, OspfSR.self->srgb);

  rc = 0;
  for (ALL_LIST_ELEMENTS_RO(srp->route->paths, node, path)) {

    osr_debug("    |-  Process new route via %pI4 for this prefix",
        &path->nexthop);

    /*
     * Get SR-Node for this nexthop. Could be not yet available
     * as Extended Link / Prefix and Router Information are flooded
     * after LSA Type 1 & 2 which populate the OSPF Route Table
     */
    srnext = get_sr_node_by_nexthop(top, path->nexthop);
    if (srnext == NULL)
      continue;

    /* And store this information for later update */
    srnext->neighbor = OspfSR.self;
    path->srni.nexthop = srnext;

    /*
     * SR Node could be known, but SRGB could be not initialize
     * This is due to the fact that Extended Link / Prefix could
     * be received before corresponding Router Information LSA
     */
    if (srnext == NULL || srnext->srgb.lower_bound == 0
        || srnext->srgb.range_size == 0) {
      osr_debug(
        "    |-  SR-Node %pI4 not ready. Stop process",
        &srnext->adv_router);
      path->srni.label_out = MPLS_INVALID_LABEL;
      continue;
    }

    osr_debug("    |-  Found SRGB %u/%u for next hop SR-Node %pI4",
        srnext->srgb.range_size, srnext->srgb.lower_bound,
        &srnext->adv_router);

    /* Compute Output Label with Nexthop SR Node SRGB */
    path->srni.label_out = sr_prefix_out_label(srp, srnext);

    osr_debug("    |-  Computed new labels in: %u out: %u",
        srp->label_in, path->srni.label_out);
    rc = 1;
  }
  return rc;
}

/* Add new NHLFE entry for Adjacency SID */
static inline void add_adj_sid(struct sr_nhlfe nhlfe)
{
  if (nhlfe.label_in != 0)
    ospf_zebra_send_adjacency_sid(ZEBRA_MPLS_LABELS_ADD, nhlfe);
}

/* Remove NHLFE entry for Adjacency SID */
static inline void del_adj_sid(struct sr_nhlfe nhlfe)
{
  if (nhlfe.label_in != 0)
    ospf_zebra_send_adjacency_sid(ZEBRA_MPLS_LABELS_DELETE, nhlfe);
}

/* Update NHLFE entry for Adjacency SID */
static inline void update_adj_sid(struct sr_nhlfe n1, struct sr_nhlfe n2)
{
  del_adj_sid(n1);
  add_adj_sid(n2);
}

/*
 * Functions to parse and get Extended Link / Prefix
 * TLVs and SubTLVs
 */

/* Extended Link SubTLVs Getter */
static struct sr_link *get_ext_link_sid(struct tlv_header *tlvh, size_t size)
{

  struct sr_link *srl;
  struct ext_tlv_link *link = (struct ext_tlv_link *)tlvh;
  struct ext_subtlv_adj_sid *adj_sid;
  struct ext_subtlv_lan_adj_sid *lan_sid;
  struct ext_subtlv_rmt_itf_addr *rmt_itf;

  struct tlv_header *sub_tlvh;
  uint16_t length = 0, sum = 0, i = 0;

  /* Check TLV size */
  if ((ntohs(tlvh->length) > size)
      || ntohs(tlvh->length) < EXT_TLV_LINK_SIZE) {
    zlog_warn("Wrong Extended Link TLV size. Abort!");
    return NULL;
  }

  srl = XCALLOC(MTYPE_OSPF_SR_PARAMS, sizeof(struct sr_link));

  /* Initialize TLV browsing */
  length = ntohs(tlvh->length) - EXT_TLV_LINK_SIZE;
  sub_tlvh = (struct tlv_header *)((char *)(tlvh) + TLV_HDR_SIZE
           + EXT_TLV_LINK_SIZE);
  for (; sum < length && sub_tlvh; sub_tlvh = TLV_HDR_NEXT(sub_tlvh)) {
    switch (ntohs(sub_tlvh->type)) {
    case EXT_SUBTLV_ADJ_SID:
      adj_sid = (struct ext_subtlv_adj_sid *)sub_tlvh;
      srl->type = ADJ_SID;
      i = CHECK_FLAG(adj_sid->flags,
               EXT_SUBTLV_LINK_ADJ_SID_BFLG)
            ? 1
            : 0;
      srl->flags[i] = adj_sid->flags;
      if (CHECK_FLAG(adj_sid->flags,
               EXT_SUBTLV_LINK_ADJ_SID_VFLG))
        srl->sid[i] = GET_LABEL(ntohl(adj_sid->value));
      else
        srl->sid[i] = ntohl(adj_sid->value);
      IPV4_ADDR_COPY(&srl->nhlfe[i].nexthop, &link->link_id);
      break;
    case EXT_SUBTLV_LAN_ADJ_SID:
      lan_sid = (struct ext_subtlv_lan_adj_sid *)sub_tlvh;
      srl->type = LAN_ADJ_SID;
      i = CHECK_FLAG(lan_sid->flags,
               EXT_SUBTLV_LINK_ADJ_SID_BFLG)
            ? 1
            : 0;
      srl->flags[i] = lan_sid->flags;
      if (CHECK_FLAG(lan_sid->flags,
               EXT_SUBTLV_LINK_ADJ_SID_VFLG))
        srl->sid[i] = GET_LABEL(ntohl(lan_sid->value));
      else
        srl->sid[i] = ntohl(lan_sid->value);
      IPV4_ADDR_COPY(&srl->nhlfe[i].nexthop,
               &lan_sid->neighbor_id);
      break;
    case EXT_SUBTLV_RMT_ITF_ADDR:
      rmt_itf = (struct ext_subtlv_rmt_itf_addr *)sub_tlvh;
      IPV4_ADDR_COPY(&srl->nhlfe[0].nexthop, &rmt_itf->value);
      IPV4_ADDR_COPY(&srl->nhlfe[1].nexthop, &rmt_itf->value);
      break;
    default:
      break;
    }
    sum += TLV_SIZE(sub_tlvh);
  }

  IPV4_ADDR_COPY(&srl->itf_addr, &link->link_data);

  osr_debug("  |-  Found primary %u and backup %u Adj/Lan Sid for %pI4",
      srl->sid[0], srl->sid[1], &srl->itf_addr);

  return srl;
}

/* Extended Prefix SubTLVs Getter */
static struct sr_prefix *get_ext_prefix_sid(struct tlv_header *tlvh,
              size_t size)
{

  struct sr_prefix *srp;
  struct ext_tlv_prefix *pref = (struct ext_tlv_prefix *)tlvh;
  struct ext_subtlv_prefix_sid *psid;

  struct tlv_header *sub_tlvh;
  uint16_t length = 0, sum = 0;

  /* Check TLV size */
  if ((ntohs(tlvh->length) > size)
      || ntohs(tlvh->length) < EXT_TLV_PREFIX_SIZE) {
    zlog_warn("Wrong Extended Link TLV size. Abort!");
    return NULL;
  }

  srp = XCALLOC(MTYPE_OSPF_SR_PARAMS, sizeof(struct sr_prefix));

  /* Initialize TLV browsing */
  length = ntohs(tlvh->length) - EXT_TLV_PREFIX_SIZE;
  sub_tlvh = (struct tlv_header *)((char *)(tlvh) + TLV_HDR_SIZE
           + EXT_TLV_PREFIX_SIZE);
  for (; sum < length && sub_tlvh; sub_tlvh = TLV_HDR_NEXT(sub_tlvh)) {
    switch (ntohs(sub_tlvh->type)) {
    case EXT_SUBTLV_PREFIX_SID:
      psid = (struct ext_subtlv_prefix_sid *)sub_tlvh;
      if (psid->algorithm != SR_ALGORITHM_SPF) {
        flog_err(EC_OSPF_INVALID_ALGORITHM,
           "SR (%s): Unsupported Algorithm",
           __func__);
        XFREE(MTYPE_OSPF_SR_PARAMS, srp);
        return NULL;
      }
      srp->type = PREF_SID;
      srp->flags = psid->flags;
      if (CHECK_FLAG(psid->flags, EXT_SUBTLV_PREFIX_SID_VFLG))
        srp->sid = GET_LABEL(ntohl(psid->value));
      else
        srp->sid = ntohl(psid->value);
      IPV4_ADDR_COPY(&srp->prefv4.prefix, &pref->address);
      srp->prefv4.prefixlen = pref->pref_length;
      srp->prefv4.family = AF_INET;
      apply_mask_ipv4(&srp->prefv4);
      break;
    default:
      break;
    }
    sum += TLV_SIZE(sub_tlvh);
  }

  osr_debug("  |-  Found SID %u for prefix %pFX", srp->sid,
      (struct prefix *)&srp->prefv4);

  return srp;
}

/*
 * Functions to manipulate Segment Routing Link & Prefix structures
 */

/* Compare two Segment Link: return 0 if equal, 1 otherwise */
static inline int sr_link_cmp(struct sr_link *srl1, struct sr_link *srl2)
{
  if ((srl1->sid[0] == srl2->sid[0]) && (srl1->sid[1] == srl2->sid[1])
      && (srl1->type == srl2->type) && (srl1->flags[0] == srl2->flags[0])
      && (srl1->flags[1] == srl2->flags[1]))
    return 0;
  else
    return 1;
}

/* Compare two Segment Prefix: return 0 if equal, 1 otherwise */
static inline int sr_prefix_cmp(struct sr_prefix *srp1, struct sr_prefix *srp2)
{
  if ((srp1->sid == srp2->sid) && (srp1->flags == srp2->flags))
    return 0;
  else
    return 1;
}

/* Update Segment Link of given Segment Routing Node */
static void update_ext_link_sid(struct sr_node *srn, struct sr_link *srl,
        uint8_t lsa_flags)
{
  struct listnode *node;
  struct sr_link *lk;
  bool found = false;
  bool config = true;

  /* Sanity check */
  if ((srn == NULL) || (srl == NULL))
    return;

  osr_debug("  |-  Process Extended Link Adj/Lan-SID");

  /* Detect if Adj/Lan_Adj SID must be configured */
  if (!CHECK_FLAG(lsa_flags, OSPF_LSA_SELF)
      && (CHECK_FLAG(srl->flags[0], EXT_SUBTLV_LINK_ADJ_SID_LFLG)
    || CHECK_FLAG(srl->flags[1], EXT_SUBTLV_LINK_ADJ_SID_LFLG)))
    config = false;

  /* Search for existing Segment Link */
  for (ALL_LIST_ELEMENTS_RO(srn->ext_link, node, lk))
    if (lk->instance == srl->instance) {
      found = true;
      break;
    }

  osr_debug("  |-  %s SR Link 8.0.0.%u for SR node %pI4",
      found ? "Update" : "Add", GET_OPAQUE_ID(srl->instance),
      &srn->adv_router);

  /* if not found, add new Segment Link and install NHLFE */
  if (!found) {
    /* Complete SR-Link and add it to SR-Node list */
    srl->srn = srn;
    IPV4_ADDR_COPY(&srl->adv_router, &srn->adv_router);
    listnode_add(srn->ext_link, srl);
    /* Try to set MPLS table */
    if (config && compute_link_nhlfe(srl)) {
      add_adj_sid(srl->nhlfe[0]);
      add_adj_sid(srl->nhlfe[1]);
    }
  } else {
    /* Update SR-Link if they are different */
    if (sr_link_cmp(lk, srl)) {
      /* Try to set MPLS table */
      if (config) {
        if (compute_link_nhlfe(srl)) {
          update_adj_sid(lk->nhlfe[0],
                   srl->nhlfe[0]);
          update_adj_sid(lk->nhlfe[1],
                   srl->nhlfe[1]);
        } else {
          del_adj_sid(lk->nhlfe[0]);
          del_adj_sid(lk->nhlfe[1]);
        }
      }
      /* Replace SR-Link in SR-Node Adjacency List */
      listnode_delete(srn->ext_link, lk);
      XFREE(MTYPE_OSPF_SR_PARAMS, lk);
      srl->srn = srn;
      IPV4_ADDR_COPY(&srl->adv_router, &srn->adv_router);
      listnode_add(srn->ext_link, srl);
    } else {
      /*
       * This is just an LSA refresh.
       * Stop processing and free SR Link
       */
      XFREE(MTYPE_OSPF_SR_PARAMS, srl);
    }
  }
}

/* Update Segment Prefix of given Segment Routing Node */
static void update_ext_prefix_sid(struct sr_node *srn, struct sr_prefix *srp)
{

  struct listnode *node;
  struct sr_prefix *pref;
  bool found = false;

  /* Sanity check */
  if (srn == NULL || srp == NULL)
    return;

  osr_debug("  |-  Process Extended Prefix SID %u", srp->sid);

  /* Process only Global Prefix SID */
  if (CHECK_FLAG(srp->flags, EXT_SUBTLV_PREFIX_SID_LFLG))
    return;

  /* Search for existing Segment Prefix */
  for (ALL_LIST_ELEMENTS_RO(srn->ext_prefix, node, pref))
    if (pref->instance == srp->instance
        && prefix_same((struct prefix *)&srp->prefv4,
           &pref->prefv4)) {
      found = true;
      break;
    }

  osr_debug("  |-  %s SR LSA ID 7.0.0.%u for SR node %pI4",
      found ? "Update" : "Add", GET_OPAQUE_ID(srp->instance),
      &srn->adv_router);

  /* Complete SR-Prefix */
  srp->srn = srn;
  IPV4_ADDR_COPY(&srp->adv_router, &srn->adv_router);

  /* if not found, add new Segment Prefix and install NHLFE */
  if (!found) {
    /* Add it to SR-Node list ... */
    listnode_add(srn->ext_prefix, srp);
    /* ... and try to set MPLS table */
    if (compute_prefix_nhlfe(srp) == 1)
      ospf_zebra_update_prefix_sid(srp);
  } else {
    /*
     * An old SR prefix exist. Check if something changes or if it
     * is just a refresh.
     */
    if (sr_prefix_cmp(pref, srp)) {
      if (compute_prefix_nhlfe(srp) == 1) {
        ospf_zebra_delete_prefix_sid(pref);
        /* Replace Segment Prefix */
        listnode_delete(srn->ext_prefix, pref);
        XFREE(MTYPE_OSPF_SR_PARAMS, pref);
        listnode_add(srn->ext_prefix, srp);
        ospf_zebra_update_prefix_sid(srp);
      } else {
        /* New NHLFE was not found.
         * Just free the SR Prefix
         */
        XFREE(MTYPE_OSPF_SR_PARAMS, srp);
      }
    } else {
      /* This is just an LSA refresh.
       * Stop processing and free SR Prefix
       */
      XFREE(MTYPE_OSPF_SR_PARAMS, srp);
    }
  }
}

/*
 * When change the FRR Self SRGB, update the NHLFE Input Label
 * for all Extended Prefix with SID index through hash_iterate()
 */
static void update_in_nhlfe(struct hash_bucket *bucket, void *args)
{
  struct listnode *node;
  struct sr_node *srn = (struct sr_node *)bucket->data;
  struct sr_prefix *srp;

  /* Process Every Extended Prefix for this SR-Node */
  for (ALL_LIST_ELEMENTS_RO(srn->ext_prefix, node, srp)) {
    /* Process Self SRN only if NO-PHP is requested */
    if ((srn == OspfSR.self)
        && !CHECK_FLAG(srp->flags, EXT_SUBTLV_PREFIX_SID_NPFLG))
      continue;

    /* Process only SID Index */
    if (CHECK_FLAG(srp->flags, EXT_SUBTLV_PREFIX_SID_VFLG))
      continue;

    /* First, remove old MPLS table entries ... */
    ospf_zebra_delete_prefix_sid(srp);
    /* ... then compute new input label ... */
    srp->label_in = index2label(srp->sid, OspfSR.self->srgb);
    /* ... and install new MPLS LFIB */
    ospf_zebra_update_prefix_sid(srp);
  }
}

/*
 * When SRGB has changed, update NHLFE Output Label for all Extended Prefix
 * with SID index which use the given SR-Node as nexthop through hash_iterate()
 */
static void update_out_nhlfe(struct hash_bucket *bucket, void *args)
{
  struct listnode *node, *pnode;
  struct sr_node *srn = (struct sr_node *)bucket->data;
  struct sr_node *srnext = (struct sr_node *)args;
  struct sr_prefix *srp;
  struct ospf_path *path;

  /* Skip Self SR-Node */
  if (srn == OspfSR.self)
    return;

  osr_debug("SR (%s): Update Out NHLFE for neighbor SR-Node %pI4",
      __func__, &srn->adv_router);

  for (ALL_LIST_ELEMENTS_RO(srn->ext_prefix, node, srp)) {
    /* Skip Prefix that has not yet a valid route */
    if (srp->route == NULL)
      continue;

    for (ALL_LIST_ELEMENTS_RO(srp->route->paths, pnode, path)) {
      /* Skip path that has not next SR-Node as nexthop */
      if (path->srni.nexthop != srnext)
        continue;

      /* Compute new Output Label */
      path->srni.label_out = sr_prefix_out_label(srp, srnext);
    }

    /* Finally update MPLS table */
    ospf_zebra_update_prefix_sid(srp);
  }
}

/*
 * Following functions are call when new Segment Routing LSA are received
 *  - Router Information: ospf_sr_ri_lsa_update() & ospf_sr_ri_lsa_delete()
 *  - Extended Link: ospf_sr_ext_link_update() & ospf_sr_ext_link_delete()
 *  - Extended Prefix: ospf_ext_prefix_update() & ospf_sr_ext_prefix_delete()
 */

/* Update Segment Routing from Router Information LSA */
void ospf_sr_ri_lsa_update(struct ospf_lsa *lsa)
{
  struct sr_node *srn;
  struct tlv_header *tlvh;
  struct lsa_header *lsah = lsa->data;
  struct ri_sr_tlv_sid_label_range *ri_srgb = NULL;
  struct ri_sr_tlv_sid_label_range *ri_srlb = NULL;
  struct ri_sr_tlv_sr_algorithm *algo = NULL;
  struct sr_block srgb;
  uint16_t length = 0, sum = 0;
  uint8_t msd = 0;

  osr_debug("SR (%s): Process Router Information LSA 4.0.0.%u from %pI4",
      __func__, GET_OPAQUE_ID(ntohl(lsah->id.s_addr)),
      &lsah->adv_router);

  /* Sanity check */
  if (IS_LSA_SELF(lsa))
    return;

  if (OspfSR.neighbors == NULL) {
    flog_err(EC_OSPF_SR_INVALID_DB,
       "SR (%s): Abort! no valid SR DataBase", __func__);
    return;
  }

  /* Search SR Node in hash table from Router ID */
  srn = (struct sr_node *)hash_lookup(OspfSR.neighbors,
              &lsah->adv_router);


  /* Collect Router Information Sub TLVs */
  /* Initialize TLV browsing */
  length = lsa->size - OSPF_LSA_HEADER_SIZE;
  srgb.range_size = 0;
  srgb.lower_bound = 0;

  for (tlvh = TLV_HDR_TOP(lsah); (sum < length) && (tlvh != NULL);
       tlvh = TLV_HDR_NEXT(tlvh)) {
    switch (ntohs(tlvh->type)) {
    case RI_SR_TLV_SR_ALGORITHM:
      algo = (struct ri_sr_tlv_sr_algorithm *)tlvh;
      break;
    case RI_SR_TLV_SRGB_LABEL_RANGE:
      ri_srgb = (struct ri_sr_tlv_sid_label_range *)tlvh;
      break;
    case RI_SR_TLV_SRLB_LABEL_RANGE:
      ri_srlb = (struct ri_sr_tlv_sid_label_range *)tlvh;
      break;
    case RI_SR_TLV_NODE_MSD:
      msd = ((struct ri_sr_tlv_node_msd *)(tlvh))->value;
      break;
    default:
      break;
    }
    sum += TLV_SIZE(tlvh);
  }

  /* Check if Segment Routing Capabilities has been found */
  if (ri_srgb == NULL) {
    /* Skip Router Information without SR capabilities
     * advertise by a non SR Node */
    if (srn == NULL) {
      return;
    } else {
      /* Remove SR Node that advertise Router Information
       * without SR capabilities. This could correspond to a
       * Node stopping Segment Routing */
      hash_release(OspfSR.neighbors, &(srn->adv_router));
      sr_node_del(srn);
      return;
    }
  }

  /* Check that RI LSA belongs to the correct SR Node */
  if ((srn != NULL) && (srn->instance != 0)
      && (srn->instance != ntohl(lsah->id.s_addr))) {
    flog_err(EC_OSPF_SR_INVALID_LSA_ID,
       "SR (%s): Abort! Wrong LSA ID 4.0.0.%u for SR node %pI4/%u",
       __func__, GET_OPAQUE_ID(ntohl(lsah->id.s_addr)),
       &lsah->adv_router, srn->instance);
    return;
  }

  /* OK. All things look good. Get SRGB */
  srgb.range_size = GET_RANGE_SIZE(ntohl(ri_srgb->size));
  srgb.lower_bound = GET_LABEL(ntohl(ri_srgb->lower.value));

  /* Check if it is a new SR Node or not */
  if (srn == NULL) {
    /* Get a new SR Node in hash table from Router ID */
    srn = (struct sr_node *)hash_get(OspfSR.neighbors,
             &lsah->adv_router,
             (void *)sr_node_new);
    /* update LSA ID */
    srn->instance = ntohl(lsah->id.s_addr);
    /* Copy SRGB */
    srn->srgb.range_size = srgb.range_size;
    srn->srgb.lower_bound = srgb.lower_bound;
  }

  /* Update Algorithm, SRLB and MSD if present */
  if (algo != NULL) {
    int i;
    for (i = 0; i < ntohs(algo->header.length); i++)
      srn->algo[i] = algo->value[0];
    for (; i < ALGORITHM_COUNT; i++)
      srn->algo[i] = SR_ALGORITHM_UNSET;
  } else {
    srn->algo[0] = SR_ALGORITHM_SPF;
  }
  srn->msd = msd;
  if (ri_srlb != NULL) {
    srn->srlb.range_size = GET_RANGE_SIZE(ntohl(ri_srlb->size));
    srn->srlb.lower_bound = GET_LABEL(ntohl(ri_srlb->lower.value));
  }

  /* Check if SRGB has changed */
  if ((srn->srgb.range_size == srgb.range_size)
      && (srn->srgb.lower_bound == srgb.lower_bound))
    return;

  /* Copy SRGB */
  srn->srgb.range_size = srgb.range_size;
  srn->srgb.lower_bound = srgb.lower_bound;

  osr_debug("  |- Update SR-Node[%pI4], SRGB[%u/%u], SRLB[%u/%u], Algo[%u], MSD[%u]",
      &srn->adv_router, srn->srgb.lower_bound, srn->srgb.range_size,
      srn->srlb.lower_bound, srn->srlb.range_size, srn->algo[0],
      srn->msd);

  /* ... and NHLFE if it is a neighbor SR node */
  if (srn->neighbor == OspfSR.self)
    hash_iterate(OspfSR.neighbors, update_out_nhlfe, srn);
}

/*
 * Delete SR Node entry in hash table information corresponding to an expired
 * Router Information LSA
 */
void ospf_sr_ri_lsa_delete(struct ospf_lsa *lsa)
{
  struct sr_node *srn;
  struct lsa_header *lsah = lsa->data;

  osr_debug("SR (%s): Remove SR node %pI4 from lsa_id 4.0.0.%u", __func__,
      &lsah->adv_router, GET_OPAQUE_ID(ntohl(lsah->id.s_addr)));

  /* Sanity check */
  if (OspfSR.neighbors == NULL) {
    flog_err(EC_OSPF_SR_INVALID_DB,
       "SR (%s): Abort! no valid SR Data Base", __func__);
    return;
  }

  /* Release Router ID entry in SRDB hash table */
  srn = hash_release(OspfSR.neighbors, &(lsah->adv_router));

  /* Sanity check */
  if (srn == NULL) {
    flog_err(EC_OSPF_SR_NODE_CREATE,
       "SR (%s): Abort! no entry in SRDB for SR Node %pI4",
       __func__, &lsah->adv_router);
    return;
  }

  if ((srn->instance != 0) && (srn->instance != ntohl(lsah->id.s_addr))) {
    flog_err(
      EC_OSPF_SR_INVALID_LSA_ID,
      "SR (%s): Abort! Wrong LSA ID 4.0.0.%u for SR node %pI4",
      __func__, GET_OPAQUE_ID(ntohl(lsah->id.s_addr)),
      &lsah->adv_router);
    return;
  }

  /* Remove SR node */
  sr_node_del(srn);
}

/* Update Segment Routing from Extended Link LSA */
void ospf_sr_ext_link_lsa_update(struct ospf_lsa *lsa)
{
  struct sr_node *srn;
  struct tlv_header *tlvh;
  struct lsa_header *lsah = lsa->data;
  struct sr_link *srl;

  int length;

  osr_debug("SR (%s): Process Extended Link LSA 8.0.0.%u from %pI4",
      __func__, GET_OPAQUE_ID(ntohl(lsah->id.s_addr)),
      &lsah->adv_router);

  /* Sanity check */
  if (OspfSR.neighbors == NULL) {
    flog_err(EC_OSPF_SR_INVALID_DB,
       "SR (%s): Abort! no valid SR DataBase", __func__);
    return;
  }

  /* Get SR Node in hash table from Router ID */
  srn = (struct sr_node *)hash_get(OspfSR.neighbors,
           (void *)&(lsah->adv_router),
           (void *)sr_node_new);

  /* Initialize TLV browsing */
  length = lsa->size - OSPF_LSA_HEADER_SIZE;
  for (tlvh = TLV_HDR_TOP(lsah); length > 0 && tlvh;
       tlvh = TLV_HDR_NEXT(tlvh)) {
    if (ntohs(tlvh->type) == EXT_TLV_LINK) {
      /* Got Extended Link information */
      srl = get_ext_link_sid(tlvh, length);
      /* Update SID if not null */
      if (srl != NULL) {
        srl->instance = ntohl(lsah->id.s_addr);
        update_ext_link_sid(srn, srl, lsa->flags);
      }
    }
    length -= TLV_SIZE(tlvh);
  }
}

/* Delete Segment Routing from Extended Link LSA */
void ospf_sr_ext_link_lsa_delete(struct ospf_lsa *lsa)
{
  struct listnode *node;
  struct sr_link *srl;
  struct sr_node *srn;
  struct lsa_header *lsah = lsa->data;
  uint32_t instance = ntohl(lsah->id.s_addr);

  osr_debug("SR (%s): Remove Extended Link LSA 8.0.0.%u from %pI4",
      __func__, GET_OPAQUE_ID(ntohl(lsah->id.s_addr)),
      &lsah->adv_router);

  /* Sanity check */
  if (OspfSR.neighbors == NULL) {
    flog_err(EC_OSPF_SR_INVALID_DB,
       "SR (%s): Abort! no valid SR DataBase", __func__);
    return;
  }

  /* Search SR Node in hash table from Router ID */
  srn = (struct sr_node *)hash_lookup(OspfSR.neighbors,
              (void *)&(lsah->adv_router));

  /*
   * SR-Node may be NULL if it has been remove previously when
   * processing Router Information LSA deletion
   */
  if (srn == NULL) {
    flog_err(EC_OSPF_SR_INVALID_DB,
       "SR (%s): Stop! no entry in SRDB for SR Node %pI4",
       __func__, &lsah->adv_router);
    return;
  }

  /* Search for corresponding Segment Link */
  for (ALL_LIST_ELEMENTS_RO(srn->ext_link, node, srl))
    if (srl->instance == instance)
      break;

  /* Remove Segment Link if found. Note that for Neighbors, only Global
   * Adj/Lan-Adj SID are stored in the SR-DB */
  if ((srl != NULL) && (srl->instance == instance)) {
    del_adj_sid(srl->nhlfe[0]);
    del_adj_sid(srl->nhlfe[1]);
    listnode_delete(srn->ext_link, srl);
    XFREE(MTYPE_OSPF_SR_PARAMS, srl);
  }
}

/* Add (LAN)Adjacency-SID from Extended Link Information */
void ospf_sr_ext_itf_add(struct ext_itf *exti)
{
  struct sr_node *srn = OspfSR.self;
  struct sr_link *srl;

  osr_debug("SR (%s): Add Extended Link LSA 8.0.0.%u from self", __func__,
      exti->instance);

  /* Sanity check */
  if (srn == NULL)
    return;

  /* Initialize new Segment Routing Link */
  srl = XCALLOC(MTYPE_OSPF_SR_PARAMS, sizeof(struct sr_link));
  srl->srn = srn;
  srl->adv_router = srn->adv_router;
  srl->itf_addr = exti->link.link_data;
  srl->instance =
    SET_OPAQUE_LSID(OPAQUE_TYPE_EXTENDED_LINK_LSA, exti->instance);
  srl->remote_id = exti->link.link_id;
  switch (exti->stype) {
  case ADJ_SID:
    srl->type = ADJ_SID;
    /* Primary information */
    srl->flags[0] = exti->adj_sid[0].flags;
    if (CHECK_FLAG(exti->adj_sid[0].flags,
             EXT_SUBTLV_LINK_ADJ_SID_VFLG))
      srl->sid[0] = GET_LABEL(ntohl(exti->adj_sid[0].value));
    else
      srl->sid[0] = ntohl(exti->adj_sid[0].value);
    if (exti->rmt_itf_addr.header.type == 0)
      srl->nhlfe[0].nexthop = exti->link.link_id;
    else
      srl->nhlfe[0].nexthop = exti->rmt_itf_addr.value;
    /* Backup Information if set */
    if (exti->adj_sid[1].header.type == 0)
      break;
    srl->flags[1] = exti->adj_sid[1].flags;
    if (CHECK_FLAG(exti->adj_sid[1].flags,
             EXT_SUBTLV_LINK_ADJ_SID_VFLG))
      srl->sid[1] = GET_LABEL(ntohl(exti->adj_sid[1].value));
    else
      srl->sid[1] = ntohl(exti->adj_sid[1].value);
    if (exti->rmt_itf_addr.header.type == 0)
      srl->nhlfe[1].nexthop = exti->link.link_id;
    else
      srl->nhlfe[1].nexthop = exti->rmt_itf_addr.value;
    break;
  case LAN_ADJ_SID:
    srl->type = LAN_ADJ_SID;
    /* Primary information */
    srl->flags[0] = exti->lan_sid[0].flags;
    if (CHECK_FLAG(exti->lan_sid[0].flags,
             EXT_SUBTLV_LINK_ADJ_SID_VFLG))
      srl->sid[0] = GET_LABEL(ntohl(exti->lan_sid[0].value));
    else
      srl->sid[0] = ntohl(exti->lan_sid[0].value);
    if (exti->rmt_itf_addr.header.type == 0)
      srl->nhlfe[0].nexthop = exti->lan_sid[0].neighbor_id;
    else
      srl->nhlfe[0].nexthop = exti->rmt_itf_addr.value;
    /* Backup Information if set */
    if (exti->lan_sid[1].header.type == 0)
      break;
    srl->flags[1] = exti->lan_sid[1].flags;
    if (CHECK_FLAG(exti->lan_sid[1].flags,
             EXT_SUBTLV_LINK_ADJ_SID_VFLG))
      srl->sid[1] = GET_LABEL(ntohl(exti->lan_sid[1].value));
    else
      srl->sid[1] = ntohl(exti->lan_sid[1].value);
    if (exti->rmt_itf_addr.header.type == 0)
      srl->nhlfe[1].nexthop = exti->lan_sid[1].neighbor_id;
    else
      srl->nhlfe[1].nexthop = exti->rmt_itf_addr.value;
    break;
  case PREF_SID:
  case LOCAL_SID:
    /* Wrong SID Type. Abort! */
    XFREE(MTYPE_OSPF_SR_PARAMS, srl);
    return;
  }

  /* Segment Routing Link is ready, update it */
  update_ext_link_sid(srn, srl, OSPF_LSA_SELF);
}

/* Delete Prefix or (LAN)Adjacency-SID from Extended Link Information */
void ospf_sr_ext_itf_delete(struct ext_itf *exti)
{
  struct listnode *node;
  struct sr_node *srn = OspfSR.self;
  struct sr_prefix *srp = NULL;
  struct sr_link *srl = NULL;
  uint32_t instance;

  osr_debug("SR (%s): Remove Extended LSA %u.0.0.%u from self",
      __func__, exti->stype == PREF_SID ? 7 : 8, exti->instance);

  /* Sanity check: SR-Node and Extended Prefix/Link list may have been
   * removed earlier when stopping OSPF or OSPF-SR */
  if (srn == NULL || srn->ext_prefix == NULL || srn->ext_link == NULL)
    return;

  if (exti->stype == PREF_SID) {
    instance = SET_OPAQUE_LSID(OPAQUE_TYPE_EXTENDED_PREFIX_LSA,
             exti->instance);
    for (ALL_LIST_ELEMENTS_RO(srn->ext_prefix, node, srp))
      if (srp->instance == instance)
        break;

    /* Uninstall Segment Prefix SID if found */
    if ((srp != NULL) && (srp->instance == instance))
      ospf_zebra_delete_prefix_sid(srp);
  } else {
    /* Search for corresponding Segment Link for self SR-Node */
    instance = SET_OPAQUE_LSID(OPAQUE_TYPE_EXTENDED_LINK_LSA,
             exti->instance);
    for (ALL_LIST_ELEMENTS_RO(srn->ext_link, node, srl))
      if (srl->instance == instance)
        break;

    /* Remove Segment Link if found */
    if ((srl != NULL) && (srl->instance == instance)) {
      del_adj_sid(srl->nhlfe[0]);
      del_adj_sid(srl->nhlfe[1]);
      listnode_delete(srn->ext_link, srl);
      XFREE(MTYPE_OSPF_SR_PARAMS, srl);
    }
  }
}

/* Update Segment Routing from Extended Prefix LSA */
void ospf_sr_ext_prefix_lsa_update(struct ospf_lsa *lsa)
{
  struct sr_node *srn;
  struct tlv_header *tlvh;
  struct lsa_header *lsah = (struct lsa_header *)lsa->data;
  struct sr_prefix *srp;

  int length;

  osr_debug("SR (%s): Process Extended Prefix LSA 7.0.0.%u from %pI4",
      __func__, GET_OPAQUE_ID(ntohl(lsah->id.s_addr)),
      &lsah->adv_router);

  /* Sanity check */
  if (OspfSR.neighbors == NULL) {
    flog_err(EC_OSPF_SR_INVALID_DB,
       "SR (%s): Abort! no valid SR DataBase", __func__);
    return;
  }

  /* Get SR Node in hash table from Router ID */
  srn = (struct sr_node *)hash_get(OspfSR.neighbors,
           (void *)&(lsah->adv_router),
           (void *)sr_node_new);
  /* Initialize TLV browsing */
  length = lsa->size - OSPF_LSA_HEADER_SIZE;
  for (tlvh = TLV_HDR_TOP(lsah); length > 0 && tlvh;
       tlvh = TLV_HDR_NEXT(tlvh)) {
    if (ntohs(tlvh->type) == EXT_TLV_LINK) {
      /* Got Extended Link information */
      srp = get_ext_prefix_sid(tlvh, length);
      /* Update SID if not null */
      if (srp != NULL) {
        srp->instance = ntohl(lsah->id.s_addr);
        update_ext_prefix_sid(srn, srp);
      }
    }
    length -= TLV_SIZE(tlvh);
  }
}

/* Delete Segment Routing from Extended Prefix LSA */
void ospf_sr_ext_prefix_lsa_delete(struct ospf_lsa *lsa)
{
  struct listnode *node;
  struct sr_prefix *srp;
  struct sr_node *srn;
  struct lsa_header *lsah = (struct lsa_header *)lsa->data;
  uint32_t instance = ntohl(lsah->id.s_addr);

  osr_debug("SR (%s): Remove Extended Prefix LSA 7.0.0.%u from %pI4",
      __func__, GET_OPAQUE_ID(ntohl(lsah->id.s_addr)),
      &lsah->adv_router);

  /* Sanity check */
  if (OspfSR.neighbors == NULL) {
    flog_err(EC_OSPF_SR_INVALID_DB,
       "SR (%s): Abort! no valid SR DataBase", __func__);
    return;
  }

  /* Search SR Node in hash table from Router ID */
  srn = (struct sr_node *)hash_lookup(OspfSR.neighbors,
              (void *)&(lsah->adv_router));

  /*
   * SR-Node may be NULL if it has been remove previously when
   * processing Router Information LSA deletion
   */
  if (srn == NULL) {
    flog_err(EC_OSPF_SR_INVALID_DB,
       "SR (%s):  Stop! no entry in SRDB for SR Node %pI4",
       __func__, &lsah->adv_router);
    return;
  }

  /* Search for corresponding Segment Prefix */
  for (ALL_LIST_ELEMENTS_RO(srn->ext_prefix, node, srp))
    if (srp->instance == instance)
      break;

  /* Remove Prefix if found */
  if ((srp != NULL) && (srp->instance == instance)) {
    ospf_zebra_delete_prefix_sid(srp);
    listnode_delete(srn->ext_prefix, srp);
    XFREE(MTYPE_OSPF_SR_PARAMS, srp);
  } else {
    flog_err(
      EC_OSPF_SR_INVALID_DB,
      "SR (%s): Didn't found corresponding SR Prefix 7.0.0.%u for SR Node %pI4",
      __func__, GET_OPAQUE_ID(ntohl(lsah->id.s_addr)),
      &lsah->adv_router);
  }
}

/*
 * Update Prefix SID. Call by ospf_ext_pref_ism_change to
 * complete initial CLI command at startup.
 *
 * @param ifp - Loopback interface
 * @param pref - Prefix address of this interface
 *
 * @return - void
 */
void ospf_sr_update_local_prefix(struct interface *ifp, struct prefix *p)
{
  struct listnode *node;
  struct sr_prefix *srp;

  /* Sanity Check */
  if ((ifp == NULL) || (p == NULL))
    return;

  /*
   * Search if there is a Segment Prefix that correspond to this
   * interface or prefix, and update it if found
   */
  for (ALL_LIST_ELEMENTS_RO(OspfSR.self->ext_prefix, node, srp)) {
    if ((srp->nhlfe.ifindex == ifp->ifindex)
        || ((IPV4_ADDR_SAME(&srp->prefv4.prefix, &p->u.prefix4))
      && (srp->prefv4.prefixlen == p->prefixlen))) {

      /* Update Interface & Prefix info */
      srp->nhlfe.ifindex = ifp->ifindex;
      IPV4_ADDR_COPY(&srp->prefv4.prefix, &p->u.prefix4);
      srp->prefv4.prefixlen = p->prefixlen;
      srp->prefv4.family = p->family;
      IPV4_ADDR_COPY(&srp->nhlfe.nexthop, &p->u.prefix4);

      /* OK. Let's Schedule Extended Prefix LSA */
      srp->instance = ospf_ext_schedule_prefix_index(
        ifp, srp->sid, &srp->prefv4, srp->flags);

      osr_debug(
        "  |-  Update Node SID %pFX - %u for self SR Node",
        (struct prefix *)&srp->prefv4, srp->sid);

      /* Install SID if NO-PHP is set and not EXPLICIT-NULL */
      if (CHECK_FLAG(srp->flags, EXT_SUBTLV_PREFIX_SID_NPFLG)
          && !CHECK_FLAG(srp->flags,
             EXT_SUBTLV_PREFIX_SID_EFLG)) {
        srp->label_in = index2label(srp->sid,
                  OspfSR.self->srgb);
        srp->nhlfe.label_out = MPLS_LABEL_IMPLICIT_NULL;
        ospf_zebra_update_prefix_sid(srp);
      }
    }
  }
}

/*
 * Following functions are used to update MPLS LFIB after a SPF run
 */

static void ospf_sr_nhlfe_update(struct hash_bucket *bucket, void *args)
{

  struct sr_node *srn = (struct sr_node *)bucket->data;
  struct listnode *node;
  struct sr_prefix *srp;
  bool old;
  int rc;

  osr_debug("  |-  Update Prefix for SR Node %pI4", &srn->adv_router);

  /* Skip Self SR Node */
  if (srn == OspfSR.self)
    return;

  /* Update Extended Prefix */
  for (ALL_LIST_ELEMENTS_RO(srn->ext_prefix, node, srp)) {

    /* Keep track of valid route */
    old = srp->route != NULL;

    /* Compute the new NHLFE */
    rc = compute_prefix_nhlfe(srp);

    /* Check computation result */
    switch (rc) {
    /* Routes are not know, remove old NHLFE if any to avoid loop */
    case -1:
      if (old)
        ospf_zebra_delete_prefix_sid(srp);
      break;
    /* Routes exist but are not ready, skip it */
    case 0:
      break;
    /* There is at least one route, update NHLFE */
    case 1:
      ospf_zebra_update_prefix_sid(srp);
      break;
    default:
      break;
    }
  }
}

void ospf_sr_update_task(struct ospf *ospf)
{

  struct timeval start_time, stop_time;

  /* Check ospf and SR status */
  if ((ospf == NULL) || (OspfSR.status != SR_UP))
    return;

  monotime(&start_time);

  osr_debug("SR (%s): Start SPF update", __func__);

  hash_iterate(OspfSR.neighbors, (void (*)(struct hash_bucket *,
             void *))ospf_sr_nhlfe_update,
         NULL);

  monotime(&stop_time);

  osr_debug("SR (%s): SPF Processing Time(usecs): %lld", __func__,
      (stop_time.tv_sec - start_time.tv_sec) * 1000000LL
        + (stop_time.tv_usec - start_time.tv_usec));
}

/*
 * --------------------------------------
 * Following are vty command functions.
 * --------------------------------------
 */

/*
 * Segment Routing Router configuration
 *
 * Must be centralize as it concerns both Extended Link/Prefix LSA
 * and Router Information LSA. Choose to call it from Extended Prefix
 * write_config() call back.
 *
 * @param vty VTY output
 *
 * @return none
 */
void ospf_sr_config_write_router(struct vty *vty)
{
  struct listnode *node;
  struct sr_prefix *srp;
  uint32_t upper;

  if (OspfSR.status == SR_UP)
    vty_out(vty, " segment-routing on\n");

  upper = OspfSR.srgb.start + OspfSR.srgb.size - 1;
  if ((OspfSR.srgb.start != DEFAULT_SRGB_LABEL)
      || (OspfSR.srgb.size != DEFAULT_SRGB_SIZE))
    vty_out(vty, " segment-routing global-block %u %u",
      OspfSR.srgb.start, upper);

  if ((OspfSR.srlb.start != DEFAULT_SRLB_LABEL) ||
      (OspfSR.srlb.end != DEFAULT_SRLB_END)) {
    if ((OspfSR.srgb.start == DEFAULT_SRGB_LABEL) &&
        (OspfSR.srgb.size == DEFAULT_SRGB_SIZE))
      vty_out(vty, " segment-routing global-block %u %u",
        OspfSR.srgb.start, upper);
    vty_out(vty, " local-block %u %u\n", OspfSR.srlb.start,
      OspfSR.srlb.end);
  } else
    vty_out(vty, "\n");

  if (OspfSR.msd != 0)
    vty_out(vty, " segment-routing node-msd %u\n", OspfSR.msd);

  if (OspfSR.self != NULL) {
    for (ALL_LIST_ELEMENTS_RO(OspfSR.self->ext_prefix, node, srp)) {
      vty_out(vty, " segment-routing prefix %pFX index %u",
        &srp->prefv4, srp->sid);
      if (CHECK_FLAG(srp->flags, EXT_SUBTLV_PREFIX_SID_EFLG))
        vty_out(vty, " explicit-null\n");
      else if (CHECK_FLAG(srp->flags,
              EXT_SUBTLV_PREFIX_SID_NPFLG))
        vty_out(vty, " no-php-flag\n");
      else
        vty_out(vty, "\n");
    }
  }
}

DEFUN(ospf_sr_enable,
       ospf_sr_enable_cmd,
       "segment-routing on",
       SR_STR
       "Enable Segment Routing\n")
{

  VTY_DECLVAR_INSTANCE_CONTEXT(ospf, ospf);

  if (OspfSR.status != SR_OFF)
    return CMD_SUCCESS;

  if (ospf->vrf_id != VRF_DEFAULT) {
    vty_out(vty,
      "Segment Routing is only supported in default VRF\n");
    return CMD_WARNING_CONFIG_FAILED;
  }

  osr_debug("SR: Segment Routing: OFF -> ON");

  /* Start Segment Routing */
  OspfSR.status = SR_ON;
  ospf_sr_start(ospf);

  return CMD_SUCCESS;
}

DEFUN (no_ospf_sr_enable,
       no_ospf_sr_enable_cmd,
       "no segment-routing [on]",
       NO_STR
       SR_STR
       "Disable Segment Routing\n")
{

  if (OspfSR.status == SR_OFF)
    return CMD_SUCCESS;

  osr_debug("SR: Segment Routing: ON -> OFF");

  /* Start by Disabling Extended Link & Prefix LSA */
  ospf_ext_update_sr(false);

  /* then, disable Router Information SR parameters */
  ospf_router_info_update_sr(false, OspfSR.self);

  /* Finally, stop Segment Routing */
  ospf_sr_stop();

  return CMD_SUCCESS;
}

static int ospf_sr_enabled(struct vty *vty)
{
  if (OspfSR.status != SR_OFF)
    return 1;

  if (vty)
    vty_out(vty, "%% OSPF SR is not turned on\n");

  return 0;
}

/* tell if two ranges [r1_lower, r1_upper] and [r2_lower,r2_upper] overlap */
static bool ranges_overlap(uint32_t r1_lower, uint32_t r1_upper,
         uint32_t r2_lower, uint32_t r2_upper)
{
  return !((r1_upper < r2_lower) || (r1_lower > r2_upper));
}


/* tell if a range is valid */
static bool sr_range_is_valid(uint32_t lower, uint32_t upper, uint32_t min_size)
{
  return (upper >= lower + min_size);
}

/**
 * Update SRGB and/or SRLB using new CLI values.
 *
 * @param gb_lower  Lower bound of the SRGB
 * @param gb_upper  Upper bound of the SRGB
 * @param lb_lower  Lower bound of the SRLB
 * @param lb_upper  Upper bound of the SRLB
 *
 * @return          0 on success, -1 otherwise
 */
static int update_sr_blocks(uint32_t gb_lower, uint32_t gb_upper,
          uint32_t lb_lower, uint32_t lb_upper)
{

  /* Check if values have changed */
  bool gb_changed, lb_changed;
  uint32_t gb_size = gb_upper - gb_lower + 1;
  uint32_t lb_size = lb_upper - lb_lower + 1;

  gb_changed =
    (OspfSR.srgb.size != gb_size || OspfSR.srgb.start != gb_lower);
  lb_changed =
    (OspfSR.srlb.end != lb_upper || OspfSR.srlb.start != lb_lower);
  if (!gb_changed && !lb_changed)
    return 0;

  /* Check if SR is correctly started i.e. Label Manager connected */
  if (OspfSR.status != SR_UP) {
    OspfSR.srgb.size = gb_size;
    OspfSR.srgb.start = gb_lower;
    OspfSR.srlb.end = lb_upper;
    OspfSR.srlb.start = lb_lower;
    return 0;
  }

  /* Release old SRGB if it has changed and is active. */
  if (gb_changed) {

    sr_global_block_delete();

    /* Set new SRGB values - but do not reserve yet (we need to
     * release the SRLB too) */
    OspfSR.srgb.size = gb_size;
    OspfSR.srgb.start = gb_lower;
    if (OspfSR.self != NULL) {
      OspfSR.self->srgb.range_size = gb_size;
      OspfSR.self->srgb.lower_bound = gb_lower;
    }
  }
  /* Release old SRLB if it has changed and reserve new block as needed.
   */
  if (lb_changed) {

    sr_local_block_delete();

    /* Set new SRLB values */
    if (sr_local_block_init(lb_lower, lb_upper) < 0) {
      ospf_sr_stop();
      return -1;
    }
    if (OspfSR.self != NULL) {
      OspfSR.self->srlb.lower_bound = lb_lower;
      OspfSR.self->srlb.range_size = lb_size;
    }
  }

  /*
   * Try to reserve the new SRGB from the Label Manger. If the
   * allocation fails, disable SR until new blocks are successfully
   * allocated.
   */
  if (gb_changed) {
    if (sr_global_block_init(OspfSR.srgb.start, OspfSR.srgb.size)
        < 0) {
      ospf_sr_stop();
      return -1;
    }
  }

  /* Update Self SR-Node */
  if (OspfSR.self != NULL) {
    /* SRGB is reserved, set Router Information parameters */
    ospf_router_info_update_sr(true, OspfSR.self);

    /* and update NHLFE entries */
    if (gb_changed)
      hash_iterate(OspfSR.neighbors,
             (void (*)(struct hash_bucket *,
                 void *))update_in_nhlfe,
             NULL);

    /* and update (LAN)-Adjacency SID */
    if (lb_changed)
      ospf_ext_link_srlb_update();
  }

  return 0;
}

DEFUN(sr_global_label_range, sr_global_label_range_cmd,
      "segment-routing global-block (16-1048575) (16-1048575) [local-block (16-1048575) (16-1048575)]",
      SR_STR
      "Segment Routing Global Block label range\n"
      "Lower-bound range in decimal (16-1048575)\n"
      "Upper-bound range in decimal (16-1048575)\n"
      "Segment Routing Local Block label range\n"
      "Lower-bound range in decimal (16-1048575)\n"
      "Upper-bound range in decimal (16-1048575)\n")
{
  uint32_t lb_upper, lb_lower;
  uint32_t gb_upper, gb_lower;
  int idx_gb_low = 2, idx_gb_up = 3;
  int idx_lb_low = 5, idx_lb_up = 6;

  /* Get lower and upper bound for mandatory global-block */
  gb_lower = strtoul(argv[idx_gb_low]->arg, NULL, 10);
  gb_upper = strtoul(argv[idx_gb_up]->arg, NULL, 10);

  /* SRLB values are taken from vtysh if there, else use the known ones */
  lb_upper = argc > idx_lb_up ? strtoul(argv[idx_lb_up]->arg, NULL, 10)
            : OspfSR.srlb.end;
  lb_lower = argc > idx_lb_low ? strtoul(argv[idx_lb_low]->arg, NULL, 10)
             : OspfSR.srlb.start;

  /* check correctness of input SRGB */
  if (!sr_range_is_valid(gb_lower, gb_upper, MIN_SRGB_SIZE)) {
    vty_out(vty, "Invalid SRGB range\n");
    return CMD_WARNING_CONFIG_FAILED;
  }

  /* check correctness of SRLB */
  if (!sr_range_is_valid(lb_lower, lb_upper, MIN_SRLB_SIZE)) {
    vty_out(vty, "Invalid SRLB range\n");
    return CMD_WARNING_CONFIG_FAILED;
  }

  /* Validate SRGB against SRLB */
  if (ranges_overlap(gb_lower, gb_upper, lb_lower, lb_upper)) {
    vty_out(vty,
      "New SR Global Block (%u/%u) conflicts with Local Block (%u/%u)\n",
      gb_lower, gb_upper, lb_lower, lb_upper);
    return CMD_WARNING_CONFIG_FAILED;
  }

  if (update_sr_blocks(gb_lower, gb_upper, lb_lower, lb_upper) < 0)
    return CMD_WARNING_CONFIG_FAILED;
  else
    return CMD_SUCCESS;
}

DEFUN(no_sr_global_label_range, no_sr_global_label_range_cmd,
      "no segment-routing global-block [(16-1048575) (16-1048575) local-block (16-1048575) (16-1048575)]",
      NO_STR SR_STR
      "Segment Routing Global Block label range\n"
      "Lower-bound range in decimal (16-1048575)\n"
      "Upper-bound range in decimal (16-1048575)\n"
      "Segment Routing Local Block label range\n"
      "Lower-bound range in decimal (16-1048575)\n"
      "Upper-bound range in decimal (16-1048575)\n")
{
  if (update_sr_blocks(DEFAULT_SRGB_LABEL, DEFAULT_SRGB_END,
           DEFAULT_SRLB_LABEL, DEFAULT_SRLB_END)
      < 0)
    return CMD_WARNING_CONFIG_FAILED;
  else
    return CMD_SUCCESS;
}

DEFUN (sr_node_msd,
       sr_node_msd_cmd,
       "segment-routing node-msd (1-16)",
       SR_STR
       "Maximum Stack Depth for this router\n"
       "Maximum number of label that could be stack (1-16)\n")
{
  uint32_t msd;
  int idx = 1;

  if (!ospf_sr_enabled(vty))
    return CMD_WARNING_CONFIG_FAILED;

  /* Get MSD */
  argv_find(argv, argc, "(1-16)", &idx);
  msd = strtoul(argv[idx]->arg, NULL, 10);
  if (msd < 1 || msd > MPLS_MAX_LABELS) {
    vty_out(vty, "MSD must be comprise between 1 and %u\n",
      MPLS_MAX_LABELS);
    return CMD_WARNING_CONFIG_FAILED;
  }

  /* Check if value has changed */
  if (OspfSR.msd == msd)
    return CMD_SUCCESS;

  /* Set this router MSD */
  OspfSR.msd = msd;
  if (OspfSR.self != NULL) {
    OspfSR.self->msd = msd;

    /* Set Router Information parameters if SR is UP */
    if (OspfSR.status == SR_UP)
      ospf_router_info_update_sr(true, OspfSR.self);
  }

  return CMD_SUCCESS;
}

DEFUN (no_sr_node_msd,
  no_sr_node_msd_cmd,
  "no segment-routing node-msd [(1-16)]",
  NO_STR
  SR_STR
  "Maximum Stack Depth for this router\n"
  "Maximum number of label that could be stack (1-16)\n")
{

  if (!ospf_sr_enabled(vty))
    return CMD_WARNING_CONFIG_FAILED;

  /* unset this router MSD */
  OspfSR.msd = 0;
  if (OspfSR.self != NULL) {
    OspfSR.self->msd = 0;

    /* Set Router Information parameters if SR is UP */
    if (OspfSR.status == SR_UP)
      ospf_router_info_update_sr(true, OspfSR.self);
  }

  return CMD_SUCCESS;
}

DEFUN (sr_prefix_sid,
       sr_prefix_sid_cmd,
       "segment-routing prefix A.B.C.D/M index (0-65535) [no-php-flag|explicit-null]",
       SR_STR
       "Prefix SID\n"
       "IPv4 Prefix as A.B.C.D/M\n"
       "SID index for this prefix in decimal (0-65535)\n"
       "Index value inside SRGB (lower_bound < index < upper_bound)\n"
       "Don't request Penultimate Hop Popping (PHP)\n"
       "Upstream neighbor must replace prefix-sid with explicit null label\n")
{
  int idx = 0;
  struct prefix p, pexist;
  uint32_t index;
  struct listnode *node;
  struct sr_prefix *srp, *exist = NULL;
  struct interface *ifp;
  bool no_php_flag = false;
  bool exp_null = false;
  bool index_in_use = false;
  uint8_t desired_flags = 0;

  if (!ospf_sr_enabled(vty))
    return CMD_WARNING_CONFIG_FAILED;

  /* Get network prefix */
  argv_find(argv, argc, "A.B.C.D/M", &idx);
  if (!str2prefix(argv[idx]->arg, &p)) {
    vty_out(vty, "Invalid prefix format %s\n", argv[idx]->arg);
    return CMD_WARNING_CONFIG_FAILED;
  }

  /* Get & verify index value */
  argv_find(argv, argc, "(0-65535)", &idx);
  index = strtoul(argv[idx]->arg, NULL, 10);
  if (index > OspfSR.srgb.size - 1) {
    vty_out(vty, "Index %u must be lower than range size %u\n",
      index, OspfSR.srgb.size);
    return CMD_WARNING_CONFIG_FAILED;
  }

  /* Get options */
  no_php_flag = argv_find(argv, argc, "no-php-flag", &idx);
  exp_null = argv_find(argv, argc, "explicit-null", &idx);

  desired_flags |= no_php_flag ? EXT_SUBTLV_PREFIX_SID_NPFLG : 0;
  desired_flags |= exp_null ? EXT_SUBTLV_PREFIX_SID_NPFLG : 0;
  desired_flags |= exp_null ? EXT_SUBTLV_PREFIX_SID_EFLG : 0;

  /* Search for an existing Prefix-SID */
  for (ALL_LIST_ELEMENTS_RO(OspfSR.self->ext_prefix, node, srp)) {
    if (prefix_same((struct prefix *)&srp->prefv4, &p))
      exist = srp;
    if (srp->sid == index) {
      index_in_use = true;
      pexist = p;
    }
  }

  /* done if prefix segment already there with same index and flags */
  if (exist && exist->sid == index && exist->flags == desired_flags)
    return CMD_SUCCESS;

  /* deny if index is already in use by a distinct prefix */
  if (!exist && index_in_use) {
    vty_out(vty, "Index %u is already used by %pFX\n", index,
      &pexist);
    return CMD_WARNING_CONFIG_FAILED;
  }

  /* First, remove old NHLFE if installed */
  if (exist && CHECK_FLAG(exist->flags, EXT_SUBTLV_PREFIX_SID_NPFLG)
      && !CHECK_FLAG(exist->flags, EXT_SUBTLV_PREFIX_SID_EFLG))
    ospf_zebra_delete_prefix_sid(exist);

  /* Create new Extended Prefix to SRDB if not found */
  if (exist == NULL) {
    srp = XCALLOC(MTYPE_OSPF_SR_PARAMS, sizeof(struct sr_prefix));
    IPV4_ADDR_COPY(&srp->prefv4.prefix, &p.u.prefix4);
    srp->prefv4.prefixlen = p.prefixlen;
    srp->prefv4.family = p.family;
    srp->sid = index;
    srp->type = LOCAL_SID;
  } else {
    /* we work on the existing SR prefix */
    srp = exist;
  }

  /* Reset labels to handle flag update */
  srp->label_in = 0;
  srp->nhlfe.label_out = 0;
  srp->sid = index;
  srp->flags = desired_flags;

  /* If NO PHP flag is present, compute NHLFE and set label */
  if (no_php_flag) {
    srp->label_in = index2label(srp->sid, OspfSR.self->srgb);
    srp->nhlfe.label_out = MPLS_LABEL_IMPLICIT_NULL;
  }

  osr_debug("SR (%s): Add new index %u to Prefix %pFX", __func__, index,
      (struct prefix *)&srp->prefv4);

  /* Get Interface and check if it is a Loopback */
  ifp = if_lookup_prefix(&p, VRF_DEFAULT);
  if (ifp == NULL) {
    /*
     * Interface could be not yet available i.e. when this
     * command is in the configuration file, OSPF is not yet
     * ready. In this case, store the prefix SID for latter
     * update of this Extended Prefix
     */
    if (exist == NULL)
      listnode_add(OspfSR.self->ext_prefix, srp);
    zlog_info(
      "Interface for prefix %pFX not found. Deferred LSA flooding",
      &p);
    return CMD_SUCCESS;
  }

  if (!if_is_loopback(ifp)) {
    vty_out(vty, "interface %s is not a Loopback\n", ifp->name);
    XFREE(MTYPE_OSPF_SR_PARAMS, srp);
    return CMD_WARNING_CONFIG_FAILED;
  }
  srp->nhlfe.ifindex = ifp->ifindex;

  /* Add SR Prefix if new */
  if (!exist)
    listnode_add(OspfSR.self->ext_prefix, srp);

  /* Update Prefix SID if SR is UP */
  if (OspfSR.status == SR_UP) {
    if (no_php_flag && !exp_null)
      ospf_zebra_update_prefix_sid(srp);
  } else
    return CMD_SUCCESS;

  /* Finally, update Extended Prefix LSA id SR is UP */
  srp->instance = ospf_ext_schedule_prefix_index(
    ifp, srp->sid, &srp->prefv4, srp->flags);
  if (srp->instance == 0) {
    vty_out(vty, "Unable to set index %u for prefix %pFX\n",
      index, &p);
    return CMD_WARNING;
  }

  return CMD_SUCCESS;
}

DEFUN (no_sr_prefix_sid,
       no_sr_prefix_sid_cmd,
       "no segment-routing prefix A.B.C.D/M [index (0-65535)|no-php-flag|explicit-null]",
       NO_STR
       SR_STR
       "Prefix SID\n"
       "IPv4 Prefix as A.B.C.D/M\n"
       "SID index for this prefix in decimal (0-65535)\n"
       "Index value inside SRGB (lower_bound < index < upper_bound)\n"
       "Don't request Penultimate Hop Popping (PHP)\n"
       "Upstream neighbor must replace prefix-sid with explicit null label\n")
{
  int idx = 0;
  struct prefix p;
  struct listnode *node;
  struct sr_prefix *srp;
  struct interface *ifp;
  bool found = false;
  int rc;

  if (!ospf_sr_enabled(vty))
    return CMD_WARNING_CONFIG_FAILED;

  if (OspfSR.status != SR_UP)
    return CMD_SUCCESS;

  /* Get network prefix */
  argv_find(argv, argc, "A.B.C.D/M", &idx);
  rc = str2prefix(argv[idx]->arg, &p);
  if (!rc) {
    vty_out(vty, "Invalid prefix format %s\n", argv[idx]->arg);
    return CMD_WARNING_CONFIG_FAILED;
  }

  /* check that the prefix is already set */
  for (ALL_LIST_ELEMENTS_RO(OspfSR.self->ext_prefix, node, srp))
    if (IPV4_ADDR_SAME(&srp->prefv4.prefix, &p.u.prefix4)
        && (srp->prefv4.prefixlen == p.prefixlen)) {
      found = true;
      break;
    }

  if (!found) {
    vty_out(vty, "Prefix %s is not found. Abort!\n",
      argv[idx]->arg);
    return CMD_WARNING_CONFIG_FAILED;
  }

  osr_debug("SR (%s): Remove Prefix %pFX with index %u", __func__,
      (struct prefix *)&srp->prefv4, srp->sid);

  /* Get Interface */
  ifp = if_lookup_by_index(srp->nhlfe.ifindex, VRF_DEFAULT);
  if (ifp == NULL) {
    vty_out(vty, "interface for prefix %s not found.\n",
      argv[idx]->arg);
    /* silently remove from list */
    listnode_delete(OspfSR.self->ext_prefix, srp);
    XFREE(MTYPE_OSPF_SR_PARAMS, srp);
    return CMD_SUCCESS;
  }

  /* Update Extended Prefix LSA */
  if (!ospf_ext_schedule_prefix_index(ifp, 0, NULL, 0)) {
    vty_out(vty, "No corresponding loopback interface. Abort!\n");
    return CMD_WARNING;
  }

  /* Delete NHLFE if NO-PHP is set and EXPLICIT NULL not set */
  if (CHECK_FLAG(srp->flags, EXT_SUBTLV_PREFIX_SID_NPFLG)
      && !CHECK_FLAG(srp->flags, EXT_SUBTLV_PREFIX_SID_EFLG))
    ospf_zebra_delete_prefix_sid(srp);

  /* OK, all is clean, remove SRP from SRDB */
  listnode_delete(OspfSR.self->ext_prefix, srp);
  XFREE(MTYPE_OSPF_SR_PARAMS, srp);

  return CMD_SUCCESS;
}


static char *sr_op2str(char *buf, size_t size, mpls_label_t label_in,
           mpls_label_t label_out)
{
  if (size < 24)
    return NULL;

  switch (label_out) {
  case MPLS_LABEL_IMPLICIT_NULL:
    snprintf(buf, size, "Pop(%u)", label_in);
    break;
  case MPLS_LABEL_IPV4_EXPLICIT_NULL:
    if (label_in == MPLS_LABEL_IPV4_EXPLICIT_NULL)
      snprintf(buf, size, "no-op.");
    else
      snprintf(buf, size, "Swap(%u, null)", label_in);
    break;
  case MPLS_INVALID_LABEL:
    snprintf(buf, size, "no-op.");
    break;
  default:
    snprintf(buf, size, "Swap(%u, %u)", label_in, label_out);
    break;
  }
  return buf;
}

static void show_sr_prefix(struct sbuf *sbuf, struct json_object *json,
         struct sr_prefix *srp)
{

  struct listnode *node;
  struct ospf_path *path;
  struct interface *itf;
  json_object *json_route = NULL, *json_obj;
  char pref[19];
  char sid[22];
  char op[32];
  char buf[PREFIX_STRLEN];
  int indent = 0;

  snprintfrr(pref, 19, "%pFX", (struct prefix *)&srp->prefv4);
  snprintf(sid, 22, "SR Pfx (idx %u)", srp->sid);
  if (json) {
    json_object_string_add(json, "prefix", pref);
    json_object_int_add(json, "sid", srp->sid);
    json_object_int_add(json, "inputLabel", srp->label_in);
  } else {
    sbuf_push(sbuf, 0, "%18s  %21s  ", pref, sid);
  }

  /* Check if it is a Local Node SID */
  if (srp->type == LOCAL_SID) {
    itf = if_lookup_by_index(srp->nhlfe.ifindex, VRF_DEFAULT);
    if (json) {
      if (!json_route) {
        json_route = json_object_new_array();
        json_object_object_add(json, "prefixRoute",
                   json_route);
      }
      json_obj = json_object_new_object();
      json_object_int_add(json_obj, "outputLabel",
              srp->nhlfe.label_out);
      json_object_string_add(json_obj, "interface",
                 itf ? itf->name : "-");
      json_object_string_addf(json_obj, "nexthop", "%pI4",
            &srp->nhlfe.nexthop);
      json_object_array_add(json_route, json_obj);
    } else {
      sbuf_push(sbuf, 0, "%20s  %9s  %15s\n",
          sr_op2str(op, 32, srp->label_in,
              srp->nhlfe.label_out),
          itf ? itf->name : "-",
          inet_ntop(AF_INET, &srp->nhlfe.nexthop,
              buf, sizeof(buf)));
    }
    return;
  }

  /* Check if we have a valid path for this prefix */
  if (srp->route == NULL) {
    if (!json) {
      sbuf_push(sbuf, 0, "\n");
    }
    return;
  }

  /* Process list of OSPF paths */
  for (ALL_LIST_ELEMENTS_RO(srp->route->paths, node, path)) {
    itf = if_lookup_by_index(path->ifindex, VRF_DEFAULT);
    if (json) {
      if (!json_route) {
        json_route = json_object_new_array();
        json_object_object_add(json, "prefixRoute",
                   json_route);
      }
      json_obj = json_object_new_object();
      json_object_int_add(json_obj, "outputLabel",
              path->srni.label_out);
      json_object_string_add(json_obj, "interface",
                 itf ? itf->name : "-");
      json_object_string_addf(json_obj, "nexthop", "%pI4",
            &path->nexthop);
      json_object_array_add(json_route, json_obj);
    } else {
      sbuf_push(sbuf, indent, "%20s  %9s  %15s\n",
          sr_op2str(op, 32, srp->label_in,
              path->srni.label_out),
          itf ? itf->name : "-",
          inet_ntop(AF_INET, &path->nexthop, buf,
              sizeof(buf)));
      /* Offset to align information for ECMP */
      indent = 43;
    }
  }
}

static void show_sr_node(struct vty *vty, struct json_object *json,
       struct sr_node *srn)
{

  struct listnode *node;
  struct sr_link *srl;
  struct sr_prefix *srp;
  struct interface *itf;
  struct sbuf sbuf;
  char pref[19];
  char sid[22];
  char op[32];
  char buf[PREFIX_STRLEN];
  uint32_t upper;
  json_object *json_node = NULL, *json_algo, *json_obj;
  json_object *json_prefix = NULL, *json_link = NULL;

  /* Sanity Check */
  if (srn == NULL)
    return;

  sbuf_init(&sbuf, NULL, 0);

  if (json) {
    json_node = json_object_new_object();
    json_object_string_addf(json_node, "routerID", "%pI4",
          &srn->adv_router);
    json_object_int_add(json_node, "srgbSize",
            srn->srgb.range_size);
    json_object_int_add(json_node, "srgbLabel",
            srn->srgb.lower_bound);
    json_object_int_add(json_node, "srlbSize",
            srn->srlb.range_size);
    json_object_int_add(json_node, "srlbLabel",
            srn->srlb.lower_bound);
    json_algo = json_object_new_array();
    json_object_object_add(json_node, "algorithms", json_algo);
    for (int i = 0; i < ALGORITHM_COUNT; i++) {
      if (srn->algo[i] == SR_ALGORITHM_UNSET)
        continue;
      json_obj = json_object_new_object();
      char tmp[2];

      snprintf(tmp, sizeof(tmp), "%u", i);
      json_object_string_add(json_obj, tmp,
                 srn->algo[i] == SR_ALGORITHM_SPF
                   ? "SPF"
                   : "S-SPF");
      json_object_array_add(json_algo, json_obj);
    }
    if (srn->msd != 0)
      json_object_int_add(json_node, "nodeMsd", srn->msd);
  } else {
    sbuf_push(&sbuf, 0, "SR-Node: %pI4", &srn->adv_router);
    upper = srn->srgb.lower_bound + srn->srgb.range_size - 1;
    sbuf_push(&sbuf, 0, "\tSRGB: [%u/%u]",
        srn->srgb.lower_bound, upper);
    upper = srn->srlb.lower_bound + srn->srlb.range_size - 1;
    sbuf_push(&sbuf, 0, "\tSRLB: [%u/%u]",
        srn->srlb.lower_bound, upper);
    sbuf_push(&sbuf, 0, "\tAlgo.(s): %s",
        srn->algo[0] == SR_ALGORITHM_SPF ? "SPF" : "S-SPF");
    for (int i = 1; i < ALGORITHM_COUNT; i++) {
      if (srn->algo[i] == SR_ALGORITHM_UNSET)
        continue;
      sbuf_push(&sbuf, 0, "/%s",
          srn->algo[i] == SR_ALGORITHM_SPF ? "SPF"
                   : "S-SPF");
    }
    if (srn->msd != 0)
      sbuf_push(&sbuf, 0, "\tMSD: %u", srn->msd);
  }

  if (!json) {
    sbuf_push(&sbuf, 0,
        "\n\n    Prefix or Link       Node or Adj. SID       Label Operation  Interface          Nexthop\n");
    sbuf_push(&sbuf, 0,
        "------------------  ---------------------  --------------------  ---------  ---------------\n");
  }
  for (ALL_LIST_ELEMENTS_RO(srn->ext_prefix, node, srp)) {
    if (json) {
      if (!json_prefix) {
        json_prefix = json_object_new_array();
        json_object_object_add(json_node,
                   "extendedPrefix",
                   json_prefix);
      }
      json_obj = json_object_new_object();
      show_sr_prefix(NULL, json_obj, srp);
      json_object_array_add(json_prefix, json_obj);
    } else {
      show_sr_prefix(&sbuf, NULL, srp);
    }
  }

  for (ALL_LIST_ELEMENTS_RO(srn->ext_link, node, srl)) {
    snprintfrr(pref, 19, "%pI4/32", &srl->itf_addr);
    snprintf(sid, 22, "SR Adj. (lbl %u)", srl->sid[0]);
    itf = if_lookup_by_index(srl->nhlfe[0].ifindex, VRF_DEFAULT);
    if (json) {
      if (!json_link) {
        json_link = json_object_new_array();
        json_object_object_add(
          json_node, "extendedLink", json_link);
      }
      /* Primary Link */
      json_obj = json_object_new_object();
      json_object_string_add(json_obj, "prefix", pref);
      json_object_int_add(json_obj, "sid", srl->sid[0]);
      json_object_int_add(json_obj, "inputLabel",
              srl->nhlfe[0].label_in);
      json_object_int_add(json_obj, "outputLabel",
              srl->nhlfe[0].label_out);
      json_object_string_add(json_obj, "interface",
                 itf ? itf->name : "-");
      json_object_string_addf(json_obj, "nexthop", "%pI4",
            &srl->nhlfe[0].nexthop);
      json_object_array_add(json_link, json_obj);
      /* Backup Link */
      json_obj = json_object_new_object();
      snprintf(sid, 22, "SR Adj. (lbl %u)", srl->sid[1]);
      json_object_string_add(json_obj, "prefix", pref);
      json_object_int_add(json_obj, "sid", srl->sid[1]);
      json_object_int_add(json_obj, "inputLabel",
              srl->nhlfe[1].label_in);
      json_object_int_add(json_obj, "outputLabel",
              srl->nhlfe[1].label_out);
      json_object_string_add(json_obj, "interface",
                 itf ? itf->name : "-");
      json_object_string_addf(json_obj, "nexthop", "%pI4",
            &srl->nhlfe[1].nexthop);
      json_object_array_add(json_link, json_obj);
    } else {
      sbuf_push(&sbuf, 0, "%18s  %21s  %20s  %9s  %15s\n",
          pref, sid,
          sr_op2str(op, 32, srl->nhlfe[0].label_in,
              srl->nhlfe[0].label_out),
          itf ? itf->name : "-",
          inet_ntop(AF_INET, &srl->nhlfe[0].nexthop,
              buf, sizeof(buf)));
      snprintf(sid, 22, "SR Adj. (lbl %u)", srl->sid[1]);
      sbuf_push(&sbuf, 0, "%18s  %21s  %20s  %9s  %15s\n",
          pref, sid,
          sr_op2str(op, 32, srl->nhlfe[1].label_in,
              srl->nhlfe[1].label_out),
          itf ? itf->name : "-",
          inet_ntop(AF_INET, &srl->nhlfe[1].nexthop,
            buf, sizeof(buf)));
    }
  }
  if (json)
    json_object_array_add(json, json_node);
  else
    vty_out(vty, "%s\n", sbuf_buf(&sbuf));

  sbuf_free(&sbuf);
}

static void show_vty_srdb(struct hash_bucket *bucket, void *args)
{
  struct vty *vty = (struct vty *)args;
  struct sr_node *srn = (struct sr_node *)bucket->data;

  show_sr_node(vty, NULL, srn);
}

static void show_json_srdb(struct hash_bucket *bucket, void *args)
{
  struct json_object *json = (struct json_object *)args;
  struct sr_node *srn = (struct sr_node *)bucket->data;

  show_sr_node(NULL, json, srn);
}

DEFUN (show_ip_opsf_srdb,
       show_ip_ospf_srdb_cmd,
       "show ip ospf database segment-routing [adv-router A.B.C.D|self-originate] [json]",
       SHOW_STR
       IP_STR
       OSPF_STR
       "Database summary\n"
       "Show Segment Routing Data Base\n"
       "Advertising SR node\n"
       "Advertising SR node ID (as an IP address)\n"
       "Self-originated SR node\n"
       JSON_STR)
{
  int idx = 0;
  struct in_addr rid;
  struct sr_node *srn;
  bool uj = use_json(argc, argv);
  json_object *json = NULL, *json_node_array = NULL;

  if (OspfSR.status == SR_OFF) {
    vty_out(vty, "Segment Routing is disabled on this router\n");
    return CMD_WARNING;
  }

  if (uj) {
    json = json_object_new_object();
    json_node_array = json_object_new_array();
    json_object_string_addf(json, "srdbID", "%pI4",
          &OspfSR.self->adv_router);
    json_object_object_add(json, "srNodes", json_node_array);
  } else {
    vty_out(vty,
      "\n\t\tOSPF Segment Routing database for ID %pI4\n\n",
      &OspfSR.self->adv_router);
  }

  if (argv_find(argv, argc, "self-originate", &idx)) {
    srn = OspfSR.self;
    show_sr_node(vty, json_node_array, srn);
    if (uj)
      vty_json(vty, json);
    return CMD_SUCCESS;
  }

  if (argv_find(argv, argc, "A.B.C.D", &idx)) {
    if (!inet_aton(argv[idx]->arg, &rid)) {
      vty_out(vty, "Specified Router ID %s is invalid\n",
        argv[idx]->arg);
      return CMD_WARNING_CONFIG_FAILED;
    }
    /* Get the SR Node from the SRDB */
    srn = (struct sr_node *)hash_lookup(OspfSR.neighbors,
                (void *)&rid);
    show_sr_node(vty, json_node_array, srn);
    if (uj)
      vty_json(vty, json);
    return CMD_SUCCESS;
  }

  /* No parameters have been provided, Iterate through all the SRDB */
  if (uj) {
    hash_iterate(OspfSR.neighbors, (void (*)(struct hash_bucket *,
               void *))show_json_srdb,
           (void *)json_node_array);
    vty_json(vty, json);
  } else {
    hash_iterate(OspfSR.neighbors, (void (*)(struct hash_bucket *,
               void *))show_vty_srdb,
           (void *)vty);
  }
  return CMD_SUCCESS;
}

/* Install new CLI commands */
void ospf_sr_register_vty(void)
{
  install_element(VIEW_NODE, &show_ip_ospf_srdb_cmd);

  install_element(OSPF_NODE, &ospf_sr_enable_cmd);
  install_element(OSPF_NODE, &no_ospf_sr_enable_cmd);
  install_element(OSPF_NODE, &sr_global_label_range_cmd);
  install_element(OSPF_NODE, &no_sr_global_label_range_cmd);
  install_element(OSPF_NODE, &sr_node_msd_cmd);
  install_element(OSPF_NODE, &no_sr_node_msd_cmd);
  install_element(OSPF_NODE, &sr_prefix_sid_cmd);
  install_element(OSPF_NODE, &no_sr_prefix_sid_cmd);
}

Coverage Report

Created: 2026-01-13 06:56