Cisco NX-OS Engineering Release 5.2(4.lisp-r8-44) 

o Fixes for LISP-Multicast

o Initial commit for for Multi-Hop

o Fixs for the following bugs
   o CSCud57012 LISP:IPV6 over ipv4 lisp and vrf instance id fails.
     o If there are no v6 locators in the locator vrf we cannot find a source
       locator.Also, if the locator is down't don't add it to itr-rlocs.
   o CSCub14348 LISP: Need to clean up "show ip lisp statistics
     o Fixed how stats are calculated for locators in cache.
   o CSCuc47166 N7k does not download PETR with lower priority to cache
     o The root cause is that while finding best priority we don't
       check whether locator is reachable or not, and as a result
       if none of the best priority locators are reachable lower
       priority locators never get downloaded to ufdm.
   o CSCue06958 Netstack Process Crash after VDC Reload
     o The root cause is that api periodic thread yields in the
       middle of the loop. What can happen is that lisp exits
       in the middle and when lisp_api_periodic gets to run again all
       the datastructures are already cleaned up. Need to add some
       sanity when tsp_sched_yield() comes back
   o CSCud57502 LISP: Multiple tracebacks with LISP URIB
     o The root cause is that we are not releasing a lisp_wrlock when
       processing ufdm acks. This causes the lock stack to get full.
   o CSCue41494 LISP:MAP-CACHE and FIB inconsistent after a vrf delete
     o If the context was deleted at some point we don't re-link the
       context_id In this case use the context_id found in the message.
   o CSCub66221 LISP: Map cache shows geo-coordinates when it's an ipv4 addr
     o When lisp recovers from pss, we don't set geo back to 0xffff.
   o CSCue24332 LISP: lisp crash after doing a  LC reload
     o Upon processing ufdm ack verify that cache and notify structures are
       interlinked. If they are not, destroy nofity structure and clear out
       cache's private data.
   o CSCue44461 LISP:GLT interface lost after unconfiguring the ip addess
     o The root cause is that GLT interface doesn't get re-hooked
       to an ip enabled interface once the current interface it's
       connected too gets de-configured. Added code to track and
       re-hook GLT interface when needed.
   o CSCub22407 LISP ESM: when removing dyneid conf, remove correspoding /32
      Null0 route
     o Changed api used for walking urib routes to the correct one
   o CSCue41260 lisp default-exclude still sending lisp multicast pim joins
     o Added a check for exclude-default in multicast APIs.
   o CSCue57888 LISP multicast: clear ip lisp map-cache not recreating the
                cache
     o Fixed the walk key while resolving the map-cache entries for
       multicast routes. Since the key wsa not initialized properly,
       sometimes the walk would not happen and the map-cache entries were 
       not being resolved.    
   o CSCue24191 LISP : Issue with Timing out of default routes from petr
     o The negative map-cache was not being updated if during the refresh
       a more specific entry is received. This has been fixed.
   o CSCue39081 Huge MTS buffer leak by pm,ip,snmp
     o Increased max MTS buffer size for MTS_SAP_LISP
   o CSCue61190 with use-petr, do native multicast for negative map-cache
                entries
     o Added check in multicast API to not use the negative map-caches with
       use-petr.
   o CSCud57531 Freetown: lisp multicast enabled will not walk existing
                routes
     o Walk multicast routes when LISP is enabled.
   o CSCue79087 LISP Mcast: send map-request mask more specific than reply
     o When a more specific static map-cache entry is being added and a
       coarser map-cache entry already exists. Then do not add the static
       map-cache if it has Map-Request option set. 
   o CSCue99239 LISP: SHA2 digest in map-register messages incorrect
     o Fixed the ordering of parameters in the SHA2 API call.
   o CSCue93583 lisp:  config not effective after switch reload
     o Start a timer to recreate GLT if the initial create fails. This caused
       the failure during LISP startup.
   o CSCue71436 LSB update from data packets incorrectly done
     o Upon getting LSB update, do a map-request to refresh the map-cache
       entry.

-----------------------------
Two new commands have been added in this release:

 [no] eid-notify <addr> key <auth-key>
      This command is used for configuring a Site-GW xTR on a first-hop router.
      When this command is configured an eid-notify message is sent to the 
      site-gw xTR upon a discovery of a host. Authentication key is 
      site-gateway specific.

 [no] eid-notify authentication-key <key>
      This command is configured on a Site-GW router. This is a configuration
      of an authentication key used by this Site-GW to authenticate EID-Notify
      messages.