* move KZ_PERMISSIONS_CACHE from defs.cfg

* add KZ_PERMISSIONS_MAX_SUBNET

* do not reload on the same iteration as load

* log the real number of records added

(cherry picked from commit 8a04441266)

* name is unique

(cherry picked from commit 9e7ab3a7b9)

(cherry picked from commit c1b9123989)

(cherry picked from commit 94c97087a1)

(cherry picked from commit 59888e0429)

(cherry picked from commit 182dc2890a)

(cherry picked from commit c137a0bc8c)

(cherry picked from commit b58e7c3a36)

(cherry picked from commit 21c7772c71)

* ip replacement only if sdp doesn't have ice candidates
* don't do ip replacement from internal by default

(cherry picked from commit 1ff059c573)

* ip replacement only if sdp doesan't have ice candidates
* don't do ip replacement from internal by default

(cherry picked from commit cc7ce736a8)

(cherry picked from commit aaf8fb3433)

(cherry picked from commit 92f25e13bd)

* a in-dialog from internal where loose_route() fails can still be routed if we have the alias for the contact

(cherry picked from commit a718d0349e)

(cherry picked from commit f567093efe)

(cherry picked from commit 830e4cd9fc)

* so that acks go to proper address

(cherry picked from commit 248f758ec5)

* send_reply checks if there's a ongoing transaction and will call proper function to send the reply

(cherry picked from commit ff12d456b5)

- Must use socket_workers instead of tcp_children for local listener so
  that the override of the tcp_children value only applies to the local
  listener

* FLAG_SESSION_PROGRESS is set when
  the media server sends a 180/183
  which is triggered after a route reply.
  if there are no endpoints to bridge
  to (dnd, faulty callflow) the return code
  from the media can be 480 if not set explicitly by
  callflow action. in that case
  we were passing it to the incoming leg
  carrier/endpoint and some carriers retry
  the call which we don't want.

initialize password variable as
variables persist in the sip process

use $vn for password which supports $null value

* append endpoint to refer if flag is set

* simpler way to trace - all

* this should NOT be static
* we should collect and auto-add new

* WITH_AUTH_TOKEN is mandatory

* failed attempt to have a keepalive based on db/script

* simplify internal redirects

* missing comma

* another missing comma