17 months agores_pjsip: Fix mwi_subscribe_replaces_unsolicited type mismatch
Richard Mudgett [Mon, 6 Aug 2018 20:37:05 +0000 (15:37 -0500)]
res_pjsip: Fix mwi_subscribe_replaces_unsolicited type mismatch


Change-Id: Iccafdd0552ea8aaed647620fb14499f1bf341843

17 months agoMerge "Create --disable-binary-modules option."
George Joseph [Wed, 29 Aug 2018 11:31:54 +0000 (06:31 -0500)]
Merge "Create --disable-binary-modules option."

17 months agoMerge "alembic: increase uri column size"
Joshua Colp [Wed, 29 Aug 2018 10:20:01 +0000 (05:20 -0500)]
Merge "alembic: increase uri column size"

17 months agoCreate --disable-binary-modules option.
Corey Farrell [Sun, 26 Aug 2018 18:18:42 +0000 (14:18 -0400)]
Create --disable-binary-modules option.

This new option can be passed for ./configure or
./tests/CI/ to prevent download/install of binary

Normally enabling the categories MENUSELECT_CODECS or MENUSELECT_RES
will result in binary modules being enabled even if the build target is
incompatible with those modules.  This includes CI scripts which enable
categories before disabling specific modules.

If more binary modules are offered in the future this will help avoid
accidentally downloading them if unwanted or incompatible.  Adding a
binary module will only require creating a new menuselect entry similar
to the existing ones, it will not be necessary to modify the CI scripts.

Change-Id: I6b1bd1c75a2e48f05b8b8a45b7a7a2d00a079166

17 months agores/res_rtp_asterisk: remove debug traces generated by an empty frame
neutrino88 [Tue, 21 Aug 2018 12:59:08 +0000 (08:59 -0400)]
res/res_rtp_asterisk: remove debug traces generated by an empty frame

The realtime text timer pops regularly and sends text frames even if
the buffer is empty. This causes a lot of unecessary debug logging.

* Made red_write() test if we need to send a frame before calling

Reported by: Emmanuel BUU
Tested by: Emmanuel BUU

Change-Id: Icf81310c3b8080b615a42060afc02ab41f9523dd

17 months agoMerge "pbx_dundi: Added IPv6 support for dundi"
Jenkins2 [Mon, 27 Aug 2018 14:38:15 +0000 (09:38 -0500)]
Merge "pbx_dundi: Added IPv6 support for dundi"

17 months agoMerge "chan_sip: improved ip:port finding of peers for non-UDP transports."
George Joseph [Mon, 27 Aug 2018 12:17:39 +0000 (07:17 -0500)]
Merge "chan_sip: improved ip:port finding of peers for non-UDP transports."

17 months agochan_sip: improved ip:port finding of peers for non-UDP transports.
Jaco Kroon [Mon, 13 Aug 2018 13:12:21 +0000 (15:12 +0200)]
chan_sip: improved ip:port finding of peers for non-UDP transports.

Also remove function peer_ipcmp_cb since it's not used (according to

Prior to b2c4e8660a9c89d07041271371151779b7ec75f6 (ASTERISK_27457)
insecure=port was the defacto standard.  That commit also prevented
insecure=port from being applied for sip/tcp or sip/tls.

Into consideration there are three sets of behaviour:

1.  "previous" - before the above commit.
2.  "current" - post above commit, pre this one.
3.  "new" - post this commit.

The problem that the above commit tried to address was guests over TCP.
It succeeded in doing that but broke transport!=udp with host!=dynamic.

This commit attempts to restore sane behaviour with respect to
transport!=udp for host!=dynamic whilst still retaining the guest users
over tcp.

It should be noted that when looking for a peer, two passes are made, the
first pass doesn't have SIP_INSECURE_PORT set for the searched-for peer,
thus looking for full matches (IP + Port), the second pass sets
SIP_INSECURE_PORT, thus expecting matches on IP only where the matched
peer allows for that (in the author's opinion:  UDP with insecure=port,
or any TCP based, non-dynamic host).

In previous behaviour there was special handling for transport=tcp|tls
whereby a peer would match during the first pass if the utilized
transport was TCP|TLS (and the peer allowed that specific transport).

This behaviour was wrong, or dubious at best.  Consider two dynamic tcp
peers, both registering from the same IP (NAT), in this case either peer
could match for connections from an IP.  It's also this behaviour that
prevented SIP guests over tcp.

The above referenced commit removed this behaviour, but kept applying
the SIP_INSECURE_PORT only to WS|WSS|UDP.  Since WS and WSS is also TCP
based, the logic here should fall into the TCP category.

This patch updates things such that the previously non-explicit (TCP
behaviour) transport test gets performed explicitly (ie, matched peer
must allow for the used transport), as well as the indeterministic
source-port nature of the TCP protocol is taken into account.  The new
match algorithm now looks like:

1.  As per previous behaviour, IP address is matched first.

2.  Explicit filter with respect to transport protocol, previous
    behaviour was semi-implied in the test for TCP pure IP match - this now
    made explicit.

3.  During first pass (without SIP_INSECURE_PORT), always match on port.

4.  If doing UDP, match if matched against peer also has
    SIP_INSECURE_PORT, else don't match.

5.  Match if not a dynamic host (for non-UDP protocols)

6.  Don't match if this is WS|WSS, or we can't trust the Contact address
    (presumably due to NAT)

7.  Match (we have a valid Contact thus if the IP matches we have no
    choice, this will likely only apply to non-NAT).

To logic-test this we need a few different scenarios.  Towards this end,
I work with a set number of peers defined in sip.conf:






Test cases for UDP:

1 - incoming UDP request from
  - previous:
    - pass 1:
      * peer1 or peer2 if from port 5060 (indeterminate, depends on peer
      * peer3 if from port 5061
      * peer5 if registered from and source port matches
    - pass 2:
      * peer3
  - current: as per previous.
  - new:
    - pass 1:
      * peer2 if from port 5060
      * peer3 if from port 5061
      * peer5 if registered from and source port matches
    - pass 2:
      * peer3

2 - incoming UDP request from
  - previous:
    - pass 1:
      * peer5 if registered from and port matches
      * peer4 if source port is 5060
    - pass 2:
      * no match (guest)
  - current: as previous.
  - new as previous (with the variation that if peer5 didn't have udp as
          allowed transport it would not match peer5 whereas previous
          and current code could).

3 - incoming UDP request from anywhere else:
  - previous:
    - pass 1:
      * peer5 if registered from that address and source port matches.
    - pass 2:
      * peer5 if insecure=port is additionally set.
      * no match (guest)
  - current - as per previous
  - new - as per previous

Test cases for TCP based transports:

4 - incoming TCP request from
  - previous:
    - pass 1 (indeterministic, depends on ordering of peers in memory):
      * peer1; or
      * peer5 if peer5 registered from (irrespective of source port); or
      * peer2 if the source port happens to be 5060; or
      * peer3 if the source port happens to be 5061.
    - pass 2: cannot happen since pass 1 will always find a peer.
  - current:
    - pass 1:
      * peer1 or peer2 if from source port 5060
      * peer3 if from source port 5060
      * peer5 if registered as and source port matches
    - pass 2:
      * no match (guest)
  - new:
    - pass 1:
      * peer 1 if from port 5060
      * peer 5 if registered and source port matches
    - pass 2:
      * peer 1

5 - incoming TCP request from
  - previous (indeterminate, depends on ordering):
    - pass 1:
      * peer4; or
      * peer5 if peer5 registered from
    - pass 2: cannot happen since pass 1 will always find a peer.
  - current:
    - pass 1:
      * peer4 if source port is 5060
      * peer5 if peer5 registered as and source port matches
    - pass 2:
      * no match (guest).
  - new:
    - pass 1:
      * peer4 if source port is 5060
      * peer5 if peer5 registered as and source port matches
    - pass 2:
      * peer4

6 - incoming TCP request from anywhere else:
  - previous:
    - pass 1:
      * peer5 if registered from that address
    - pass 2: cannot happen since pass 1 will always find a peer.
  - current:
    - pass 1:
      * peer5 if registered from that address and port matches.
    - pass 2:
      * no match (guest)
  - new: as per current.

It should be noted the test cases don't make explicit mention of TLS, WS
or WSS.  WS and WSS previously followed UDP semantics, they will now
enforce source port matching.  TLS follow TCP semantics.

The previous commit specifically tried to address test-case 6, but broke
test-cases 4 and 5 in the process.

ASTERISK-27881 #close

Change-Id: I61a9804e4feba9c7224c481f7a10bf7eb7c7f2a2

17 months agoAMI: be less verbose when adding HTTP headers to AMI/HTTP messages.
Jaco Kroon [Mon, 20 Aug 2018 12:23:38 +0000 (14:23 +0200)]
AMI: be less verbose when adding HTTP headers to AMI/HTTP messages.

All HTTP/AMI message headers are being sent to the verbose channel.
There are multiple places this is happening.  Consolidate the loop into
a function.  Drop the debug/verbose message.

Convert to using ast_asprintf to perform the length calculation, memory
allocation and snprintf all in one step.

Change-Id: Ic45e673fde05bd544be95ad5cdbc69518207c1a1

17 months agoMerge "sample_configs: noload by default"
Jenkins2 [Thu, 23 Aug 2018 13:55:44 +0000 (08:55 -0500)]
Merge "sample_configs: noload by default"

17 months agoalembic: increase uri column size
Florian Floimair [Thu, 23 Aug 2018 11:57:31 +0000 (13:57 +0200)]
alembic: increase uri column size

When mobile SIP clients register with Asterisk that use some sort of
push notifications, the URI can get quite lengthy due to the
additional push-service annotations (things like tokens, pn-type, etc.)
contained in it.

ASTERISK-28022 #close

Change-Id: I4c7ceadc3bb405f3daf722641c8cd5ca4188cc37

17 months agosample_configs: noload by default
Matthew Fredrickson [Wed, 22 Aug 2018 15:50:55 +0000 (10:50 -0500)]
sample_configs: noload by default

Change disables loading of in default installation.  Loading
res_hep has a performance impact whether it's used or not.  This disables
loading of it in sample config files.

Change-Id: I5ec150cf941634fabc72973e5bf1a965cb0ef9d0

17 months agoMerge "res_pjsip: Reduce processing when a Contact is updated."
Joshua Colp [Wed, 22 Aug 2018 17:42:46 +0000 (12:42 -0500)]
Merge "res_pjsip: Reduce processing when a Contact is updated."

17 months agoapp_queue: Silence GCC 8 compiler warning
Sean Bright [Tue, 21 Aug 2018 18:50:33 +0000 (14:50 -0400)]
app_queue: Silence GCC 8 compiler warning

I'm only seeing an error in 14+, so I assume it is due to different
compiler options:

app_queue.c: In function ‘handle_queue_add_member’:
app_queue.c:10234:19: error: ‘%d’ directive writing between 1 and 11
    bytes into a region of size 3 [-Werror=format-overflow=]
     sprintf(num, "%d", state);
app_queue.c:10234:18: note: directive argument in the range
    [-2147483648, 99]
     sprintf(num, "%d", state);

Compiler: gcc version 8.0.1 20180414 (experimental)
    [trunk revision 259383] (Ubuntu 8-20180414-1ubuntu2)

Change-Id: I18577590da46829c1ea7d8b82e41d69f105baa10

17 months agoMerge "AMI: Remove docs for nonexistent AMI ContactStatus event headers"
Joshua Colp [Tue, 21 Aug 2018 23:53:11 +0000 (18:53 -0500)]
Merge "AMI: Remove docs for nonexistent AMI ContactStatus event headers"

17 months agoMerge "pbx_dundi: Fix debug frame decode string."
Joshua Colp [Tue, 21 Aug 2018 23:52:46 +0000 (18:52 -0500)]
Merge "pbx_dundi: Fix debug frame decode string."

17 months agoMerge "pbx_dundi.c: Handle thread shutdown better."
George Joseph [Tue, 21 Aug 2018 12:26:01 +0000 (07:26 -0500)]
Merge "pbx_dundi.c: Handle thread shutdown better."

17 months agoMerge "pbx_dundi.c: Misc memory management fixes when destroying peers"
Joshua Colp [Tue, 21 Aug 2018 11:27:23 +0000 (06:27 -0500)]
Merge "pbx_dundi.c: Misc memory management fixes when destroying peers"

17 months agoAMI: Remove docs for nonexistent AMI ContactStatus event headers
Richard Mudgett [Mon, 20 Aug 2018 16:23:21 +0000 (11:23 -0500)]
AMI: Remove docs for nonexistent AMI ContactStatus event headers

Change-Id: I5736965c64c44338f7330e85a24bb46818607f19

17 months agoMerge "res_rtp_asterisk.c: Fix unused variable warnings"
George Joseph [Mon, 20 Aug 2018 16:31:20 +0000 (11:31 -0500)]
Merge "res_rtp_asterisk.c: Fix unused variable warnings"

17 months agoMerge "res_sorcery_realtime.c: Fix unqualified fetch warning."
George Joseph [Mon, 20 Aug 2018 15:57:05 +0000 (10:57 -0500)]
Merge "res_sorcery_realtime.c: Fix unqualified fetch warning."

17 months agoMerge "res_pjsip_t38.c: Fix crash if already saw a final T.38 reINVITE response."
George Joseph [Mon, 20 Aug 2018 15:55:01 +0000 (10:55 -0500)]
Merge "res_pjsip_t38.c: Fix crash if already saw a final T.38 reINVITE response."

17 months agores_pjsip: Reduce processing when a Contact is updated.
Joshua Colp [Mon, 6 Aug 2018 11:22:22 +0000 (11:22 +0000)]
res_pjsip: Reduce processing when a Contact is updated.

When a Contact is updated the only material change that qualify
support cares about is the underlying configuration for the AOR.
In this case we will update things with the new AOR information but
otherwise the callback to indicate the Contact has changed can be

This is because it is only when a Contact is added or deleted that
material changes occur within the qualify support. An update can't
change the URI since it would result in a new Contact so it can be

Change-Id: I2f97ebfa79969a36a97bb7b9afd5b6268cf1a07d

18 months agores_pjsip_t38.c: Fix crash if already saw a final T.38 reINVITE response.
Richard Mudgett [Sat, 11 Aug 2018 00:28:45 +0000 (19:28 -0500)]
res_pjsip_t38.c: Fix crash if already saw a final T.38 reINVITE response.

We were still getting crashes after the first fix.  Somehow we receive a
non-2xx final response before we get a 200 final response.  With the
failure response we had already cleaned up and destroyed some data
structures.  When the unexpected 200 response comes in we crash.

* Add protection code to prevent processing another final T.38 reINVITE


Change-Id: I8b5baba8d07fe4d63f0d7d05d3eb9a3d27d40a74

18 months agores_sorcery_realtime.c: Fix unqualified fetch warning.
Richard Mudgett [Thu, 9 Aug 2018 23:46:19 +0000 (18:46 -0500)]
res_sorcery_realtime.c: Fix unqualified fetch warning.

The allow_unqualified_fetch option for the sorcery realtime backend
blocked actually fetching all rows when the option is set to warn.

* Made issue a warning and actually do the request when
allow_unqualified_fetch=warn is set.

Change-Id: I74456c80a03a62dce66fc3dc3cb0cf2351ac4312

18 months agopbx_dundi: Added IPv6 support for dundi
Kirsty Tyerman [Mon, 11 Jun 2018 05:07:17 +0000 (15:07 +1000)]
pbx_dundi: Added IPv6 support for dundi

Change includes move to netsock2 library.

Reported-by: Adam Secombe

Change-Id: Ia9e8dc3d153de7a291dbda4bd87fc827dd2bb846

18 months agopbx_dundi.c: Misc memory management fixes when destroying peers
Richard Mudgett [Thu, 16 Aug 2018 02:31:45 +0000 (21:31 -0500)]
pbx_dundi.c: Misc memory management fixes when destroying peers

* In destroy_peer(), fixed memory leaks of lookup history strings and
qualify transactions when destroying peers.

* In destroy_peer(), fixed leaving the registerexpire scheduled callback
active when a peer is destroyed on a reload.  The reload marks and sweeps
peers so any peers not explicitly configured get destroyed.  Peers created
dynamically from the '*' peer will not exist until they re-register after
the reload.  These destroyed peers caused memory corruption when the
registerexpire timer expired.

* Made build_peer() not schedule any callbacks on the '*' peer
(empty_eid).  It is a special peer that is cloned to dynamically created
peers so it doesn't actually get involved in any message transactions.

* Made do_register_expire() remove the dundi/dpeers AstDB entry when a
peer registration expires.

* Fix deep_copy_peer() to not copy some things that cannot be copied to
the cloned peer structure.  Timers, message transactions, and lookup
history are specific to a peer instance.

* Made set_config() lock around processing the mappings configuration.

* Reordered unload_module() to handle load_module() declining the load due
to error.

Change-Id: Ib846b2b60d027f3a2c2b3b563d9a83a357dce1d6

18 months agopbx_dundi.c: Handle thread shutdown better.
Richard Mudgett [Thu, 16 Aug 2018 04:49:19 +0000 (23:49 -0500)]
pbx_dundi.c: Handle thread shutdown better.

Change-Id: Id52f99bd6a948fe6dd82acc0a28b2447a224fe87

18 months agopbx_dundi: Fix debug frame decode string.
Richard Mudgett [Wed, 15 Aug 2018 23:14:52 +0000 (18:14 -0500)]
pbx_dundi: Fix debug frame decode string.

* Fixed a typo in the name of the REGREQ frame decode string array.
* Fixed off by one range check indexing into the frame decode string
* Removed some unneeded casts associated with the decode string array.

Change-Id: I77435e81cd284bab6209d545919bf236ad7933c2

18 months agopbx_dundi: Update sample config documentation.
Richard Mudgett [Thu, 16 Aug 2018 21:21:07 +0000 (16:21 -0500)]
pbx_dundi: Update sample config documentation.

Change-Id: I33d0ad0611c2124ca3440f0f811fa0f45e4e2849

18 months agores_rtp_asterisk.c: Fix unused variable warnings
Richard Mudgett [Wed, 15 Aug 2018 19:44:48 +0000 (14:44 -0500)]
res_rtp_asterisk.c: Fix unused variable warnings

Compiling without SRTP support installed resulted in some unused variable
warnings.  These warnings also showed that the srtp variable was obtained
and passed around some functions but not really used even when a system
has SRTP installed.

Change-Id: I6daad34be3e89b19adef6e2fbe738018975155fc

18 months agoMerge "res_resolver_unbound: Fix leak of config nameserver strings."
Joshua Colp [Fri, 17 Aug 2018 10:40:01 +0000 (05:40 -0500)]
Merge "res_resolver_unbound: Fix leak of config nameserver strings."

18 months agoMerge "res_pjsip: Resolve transport management leak at shutdown."
Joshua Colp [Fri, 17 Aug 2018 10:38:56 +0000 (05:38 -0500)]
Merge "res_pjsip: Resolve transport management leak at shutdown."

18 months agoMerge "res_odbc: Allow unload at shutdown."
Kevin Harwell [Thu, 16 Aug 2018 22:48:01 +0000 (17:48 -0500)]
Merge "res_odbc: Allow unload at shutdown."

18 months agoCI: Fixup for non-13 branches
George Joseph [Thu, 16 Aug 2018 18:51:51 +0000 (12:51 -0600)]
CI: Fixup for non-13 branches

Change-Id: I5e1d4a09e58b92b541bc8ed6f9e10e54c4e5101f

18 months agoCI: Final version of setting correct gerrit creds
George Joseph [Thu, 16 Aug 2018 18:28:03 +0000 (12:28 -0600)]
CI:  Final version of setting correct gerrit creds

Change-Id: I7729ecceedceb12f52bf18dae259846aa1d993b3

18 months agoCI: Add https credentials to gerrit checkouts
George Joseph [Thu, 16 Aug 2018 17:08:21 +0000 (11:08 -0600)]
CI:  Add https credentials to gerrit checkouts

If the review to be tested is in a project with restricted access,
we need to use the jenkins user's gerrit https credentials when we
do the checkout or the checkout will fail.

Change-Id: I9dc9994763c5ebfeb9f1cff60fb53f6902b7fd5f

18 months agoMerge "res/res_pjsip_sdp_rtp: put rtcp-mux in answer only if offered"
George Joseph [Thu, 16 Aug 2018 14:45:33 +0000 (09:45 -0500)]
Merge "res/res_pjsip_sdp_rtp:  put rtcp-mux in answer only if offered"

18 months agores/res_pjsip_sdp_rtp: put rtcp-mux in answer only if offered
Torrey Searle [Thu, 9 Aug 2018 07:34:17 +0000 (09:34 +0200)]
res/res_pjsip_sdp_rtp:  put rtcp-mux in answer only if offered

If in the initial sdp the caller doesn't include the line

Then asterisk shoud not include rtcp-mux in the response regardless
of rtcp-mux being enabled on the endpoint

ASTERISK-28007 #close

Change-Id: I58e9b9f40a139afc0da5de41906cc608fb62adc7

18 months agores_resolver_unbound: Fix leak of config nameserver strings.
Corey Farrell [Wed, 15 Aug 2018 19:49:01 +0000 (15:49 -0400)]
res_resolver_unbound: Fix leak of config nameserver strings.

Change-Id: I3f396316bb40d1ae6e91f5f688042420f1a540ed

18 months agores_pjsip: Resolve transport management leak at shutdown.
Corey Farrell [Wed, 15 Aug 2018 18:51:36 +0000 (14:51 -0400)]
res_pjsip: Resolve transport management leak at shutdown.

Cleanup idle check scheduled events at shutdown.

Change-Id: I61bfbb56bac69fe840c3242927d31ff3593be461

18 months agores_odbc: Allow unload at shutdown.
Corey Farrell [Wed, 15 Aug 2018 16:31:00 +0000 (12:31 -0400)]
res_odbc: Allow unload at shutdown.

This makes it possible for REF_DEBUG to report no leaks when loading

Change-Id: I1a3dea786bd6e7f4820a6dd5cbaa197fa783ce93

18 months agores_pjsip: Fix leak in pjsip_options.
Corey Farrell [Wed, 15 Aug 2018 16:12:49 +0000 (12:12 -0400)]
res_pjsip: Fix leak in pjsip_options.

sip_options_get_endpoint_state_compositor_state leaked a reference to
the first available endpoint state compositor that was found.

Change-Id: Idb6be19f7219b6eed1dfb19c1e740dd40cb3fdc7

18 months agoMerge "res_pjsip_caller_id: Add "party" parameter to RPID header."
George Joseph [Wed, 15 Aug 2018 14:44:43 +0000 (09:44 -0500)]
Merge "res_pjsip_caller_id: Add "party" parameter to RPID header."

18 months agoMerge "res_pjsip/rtp: No joint capabilities between streams."
Jenkins2 [Wed, 15 Aug 2018 14:38:12 +0000 (09:38 -0500)]
Merge "res_pjsip/rtp: No joint capabilities between streams."

18 months agoMerge "contrib/scripts: Make astgenkey executable"
George Joseph [Wed, 15 Aug 2018 12:50:48 +0000 (07:50 -0500)]
Merge "contrib/scripts: Make astgenkey executable"

18 months agoMerge "Build System: Improve ccache matching for different menuselect options."
Jenkins2 [Tue, 14 Aug 2018 18:41:32 +0000 (13:41 -0500)]
Merge "Build System: Improve ccache matching for different menuselect options."

18 months agocontrib/scripts: Make astgenkey executable
Richard Mudgett [Tue, 14 Aug 2018 16:55:42 +0000 (11:55 -0500)]
contrib/scripts: Make astgenkey executable

Change-Id: I11641d65592536dea9cbca5aa94a24c25d24dd5f

18 months agores_pjsip_caller_id: Add "party" parameter to RPID header.
Joshua Colp [Tue, 14 Aug 2018 12:29:18 +0000 (09:29 -0300)]
res_pjsip_caller_id: Add "party" parameter to RPID header.

This change adds the "party" parameter to the Remote-Party-ID header
which indicates which party the header information is applicable
to. In Asterisk this is determined on whether we are the calling
or called party. This is added to improve interoperability with some


Change-Id: I1eec3e377ffff8633b5c1dd59a05e9533122cfca

18 months agoMerge "CI: Add support for coverage processing."
Jenkins2 [Tue, 14 Aug 2018 12:34:13 +0000 (07:34 -0500)]
Merge "CI: Add support for coverage processing."

18 months agores_pjsip/rtp: No joint capabilities between streams.
Ben Ford [Tue, 7 Aug 2018 15:57:29 +0000 (10:57 -0500)]
res_pjsip/rtp: No joint capabilities between streams.

When a conference contained a mixture of audio/video and audio-only
users, a NOTICE message would pop up stating there are no joint
capabilities between streams. This happens because streams can never be
removed, but they can be in a REMOVED state. If we have the scenario
where user A joins with audio/video, user B joins with audio-only, and
user C joins with audio/video, then user A leaves, the message would
be triggered. That removed stream is still in the SDP, but Asterisk
would pass it through, causing it to be seen as a ulaw stream. A check
has been added for removed streams, setting their status to REMOVED when
handling negotiated SDPs.

Also addressed an issue where user A joins, then user B joins but does
not receive video until much later. Full frames were not being sent,
causing some PLI from the browser. Because the video was flowing in one
direction, the browser sets the SSRC to 1, but Asterisk was dropping the
PLI because of that. Added a check to see if the SSRC is 1 or not, which
sends full frames and allows video to flow between user A and user B.
This should only happen when dealing with PSFB or FUR, and in the case
of PSFB, only for PLI.


Change-Id: I26e7c6f101bc119549eeca406b5bcd25ad8ebc5e

18 months agoapp_queue: set QUEUESTATUS to LEAVEEMPTY instead of CONTINUE
Ivan Poddubny [Sun, 12 Aug 2018 16:04:42 +0000 (18:04 +0200)]
app_queue: set QUEUESTATUS to LEAVEEMPTY instead of CONTINUE

When a call leaves a queue on leaveempty condition, QUEUESTATUS
must be set to LEAVEEMPTY, no matter whether Queue was executed with or
without the "c" (continue) option.

The regression was introduced in the fix for ASTERISK_25665.
The following fix (ASTERISK_27065) was incomplete, as QUEUESTATUS was
overwritten in case when "c" is set, regardless of what was the cause
for leaving the queue.

ASTERISK-27973 #close
Reported-by: Valentin Safonov

Change-Id: Iec013fe6a26a4e825ca572a1dda4f3cee5f6f80c

18 months agoSample configs: Fix pjsip.conf syntax error.
Corey Farrell [Thu, 9 Aug 2018 20:25:41 +0000 (16:25 -0400)]
Sample configs: Fix pjsip.conf syntax error.

It is valid for a config file to be empty or contain only comments, but
not valid for a config value to be set when no uncommented context
exists.  This caused an error to be loged numerous times during start
when loading the default pjsip.conf.

Change-Id: Icf3b0d69b4ecb6e935eecd43c99ed8b32a5a1cf6

18 months agoMerge "res_pjsip_registrar: Improve performance on inbound handling."
Joshua Colp [Wed, 8 Aug 2018 17:08:49 +0000 (12:08 -0500)]
Merge "res_pjsip_registrar: Improve performance on inbound handling."

18 months agoCI: Add support for coverage processing.
Corey Farrell [Fri, 20 Jul 2018 03:28:14 +0000 (23:28 -0400)]
CI: Add support for coverage processing.

Enable coverage with `./tests/CI/ --coverage`.  This
will cause Asterisk to be compiled with coverage support.  It also
initializes 'before' coverage data for all sources.  Accept
--tested-only to disable modules which are not run by any test.
Enabling coverage also sets tested-only true by default.  To build
everything with coverage enabled use `--coverage --tested-only=0`.

./tests/CI/ is used to process the coverage and
generate HTML reports.

Fix utils/check_expr2 which failed to compiled with coverage enabled.

Add status output 5 times per stage of astobj2_test_perf to ensure
remote CLI does not timeout when compiled with coverage.  Remote CLI
disconnects if no output is received for 60 seconds.  When coverage is
enabled it takes about 70 seconds for my laptop to run the stages of
this test, so with the change a message is printed every 14 seconds.

Change-Id: I890f7d5665087426ad7d3e363187691b9afc2222

18 months agoMerge "stasis: Reduce calculation of stasis message type hash."
Joshua Colp [Wed, 8 Aug 2018 10:54:02 +0000 (05:54 -0500)]
Merge "stasis: Reduce calculation of stasis message type hash."

18 months agoMerge "res_pjsip: Make pjlib.h consistently included."
Joshua Colp [Wed, 8 Aug 2018 10:53:53 +0000 (05:53 -0500)]
Merge "res_pjsip: Make pjlib.h consistently included."

18 months agoMerge "res_pjsip.h: Fix doxygen comments."
Joshua Colp [Wed, 8 Aug 2018 10:18:44 +0000 (05:18 -0500)]
Merge "res_pjsip.h: Fix doxygen comments."

18 months agoMerge "pjproject_bundled: Fix for Solaris builds. Do not undef s_addr."
Joshua Colp [Wed, 8 Aug 2018 10:10:32 +0000 (05:10 -0500)]
Merge "pjproject_bundled: Fix for Solaris builds. Do not undef s_addr."

18 months agores_pjsip.h: Fix doxygen comments.
Richard Mudgett [Mon, 6 Aug 2018 17:19:12 +0000 (12:19 -0500)]
res_pjsip.h: Fix doxygen comments.

Change-Id: I9cf97bdc756012d1f552ab007f4aa85e0ddb4e62

18 months agostasis: Reduce calculation of stasis message type hash.
Joshua Colp [Mon, 6 Aug 2018 11:36:22 +0000 (08:36 -0300)]
stasis: Reduce calculation of stasis message type hash.

When the stasis cache is used a hash is calculated for
retrieving or inserting messages. This change calculates
a hash when the message type is initialized that is then
used each time needed. This ensures that the hash is
calculated only once for the message type.

Change-Id: I4fe6bfdafb55bf5c322dd313fbd8c32cce73ef37

18 months agoMerge "dialplan_functions: wrong srtp use status report of a dialplan function"
Joshua Colp [Mon, 6 Aug 2018 13:34:20 +0000 (08:34 -0500)]
Merge "dialplan_functions: wrong srtp use status report of a dialplan function"

18 months agoMerge "pjproject_bundled: Find shared libraries in root --with-ssl=PATH."
Joshua Colp [Mon, 6 Aug 2018 10:28:47 +0000 (05:28 -0500)]
Merge "pjproject_bundled: Find shared libraries in root --with-ssl=PATH."

18 months agopjproject_bundled: Fix for Solaris builds. Do not undef s_addr.
Alexander Traud [Mon, 30 Jul 2018 12:49:08 +0000 (14:49 +0200)]
pjproject_bundled: Fix for Solaris builds. Do not undef s_addr.

The authors of PJProject undef s_addr because of some issue in Microsoft
Windows. However in Oracle Solaris, s_addr is not a structure member, but
defined to map to the real structure member.

Updates the patch from ASTERISK_20366


Change-Id: I8223026d4d54e2a46521085fcc94bfa6ebe35b11

18 months agores_pjsip: Make pjlib.h consistently included.
Richard Mudgett [Fri, 3 Aug 2018 20:59:06 +0000 (15:59 -0500)]
res_pjsip: Make pjlib.h consistently included.

* Don't include pjlib.h twice in res_pjsip.h
* Consistently use #include <> form for pjproject includes.
(pjsip.h and pjlib.h)

Change-Id: I3f7b42044840de64edf7e9d7695cb60c45990dc7

18 months agodialplan_functions: wrong srtp use status report of a dialplan function
Salah Ahmed [Thu, 2 Aug 2018 19:37:16 +0000 (14:37 -0500)]
dialplan_functions: wrong srtp use status report of a dialplan function

If asterisk offer an endpoint with SRTP and that endpoint respond
with non srtp, in that case channel(rtp,secure,audio) reply wrong

Why delete flag AST_SRTP_CRYPTO_OFFER_OK while check identical remote_key:
Currently this flag has being set redundantly. In either case identical
or different remote_key this flag has being set. So if we
don't set it while we receive identical remote_key or non SRTP SDP
response then we can take decision of srtp use by using that flag.


Change-Id: I29dc2843cf4e5ae2604301cb4ff258f1822dc2d7

18 months agoMerge "BuildSystem: Enable ncurses for menuselect in Solaris 11."
Kevin Harwell [Fri, 3 Aug 2018 18:29:11 +0000 (13:29 -0500)]
Merge "BuildSystem: Enable ncurses for menuselect in Solaris 11."

18 months agoMerge "pjsip_wizard.conf.sample: Update remote_hosts description."
Kevin Harwell [Fri, 3 Aug 2018 18:27:24 +0000 (13:27 -0500)]
Merge "pjsip_wizard.conf.sample: Update remote_hosts description."

18 months agoMerge "res_pjsip_endpoint_identifier_ip.c: Added regex support to match_header"
Kevin Harwell [Fri, 3 Aug 2018 18:26:30 +0000 (13:26 -0500)]
Merge "res_pjsip_endpoint_identifier_ip.c: Added regex support to match_header"

18 months agopjproject_bundled: Find shared libraries in root --with-ssl=PATH.
Alexander Traud [Mon, 30 Jul 2018 11:05:34 +0000 (13:05 +0200)]
pjproject_bundled: Find shared libraries in root --with-ssl=PATH.

The script configure from Teluu expects shared libraries (.so) in a subfolder
called 'lib', when --with-xyz=PATH is specified. However for OpenSSL, the
default location is the root of the source folder = PATH. Furthermore, Asterisk
supports both, 'lib' and root. For consistency and because Asterisk is using
(only) OpenSSL in PJProject, it is enhanced to support both locations, just
like Asterisk.


Change-Id: I8eb916a88b6b8c22e29bb40bee8faaca6c73406f

18 months agores_pjsip_registrar: Improve performance on inbound handling.
Joshua Colp [Wed, 1 Aug 2018 14:45:04 +0000 (14:45 +0000)]
res_pjsip_registrar: Improve performance on inbound handling.

This change removes a sorcery lookup for retrieving all
contacts at the end of the registration process by keeping
track of the contacts that are added/updated/deleted.

This ensures at the end of the process the container of
contacts we have is the current state.

Pool usage has also been reduced by allocating one for
usage throughout the handling of a REGISTER and resetting
it to a clean state. This ensures that in most cases
we allocate once and just reuse it.


Change-Id: I1a78b2d46f9a2045dbbff1a3fd6dba84b612b3cb

18 months agoMerge "thirdparty/pjproject: fix deadlock in response retransmissions"
Joshua Colp [Thu, 2 Aug 2018 12:12:35 +0000 (07:12 -0500)]
Merge "thirdparty/pjproject: fix deadlock in response retransmissions"

18 months agoMerge "BuildSystem: Enable Jansson in Solaris 11."
Joshua Colp [Thu, 2 Aug 2018 11:39:07 +0000 (06:39 -0500)]
Merge "BuildSystem: Enable Jansson in Solaris 11."

18 months agothirdparty/pjproject: fix deadlock in response retransmissions
Torrey Searle [Tue, 17 Jul 2018 12:13:43 +0000 (14:13 +0200)]
thirdparty/pjproject: fix deadlock in response retransmissions

The tdata containing the response can be shared by both the dialog
object and the tsx object.  In order to prevent the race condition
between the tsx retransmission and the dialog sending a response,
clone the tdata before modifying it for the dialog send response.

ASTERISK-27966 #close

Change-Id: Ic381004a3a212fe1d8eca0e707fe09dba4a6ab4e

18 months agoBuild System: Improve ccache matching for different menuselect options.
Corey Farrell [Wed, 1 Aug 2018 04:54:11 +0000 (00:54 -0400)]
Build System: Improve ccache matching for different menuselect options.

Changing any Menuselect option in the `Compiler Flags` section causes a
full rebuild of the Asterisk source tree.  Every enabled option causes
a #define to be added to buildopts.h, thus breaking ccache caching for
every source file that includes "asterisk.h".  In most cases each option
only applies to one or two files.  Now we only define those options for
the specific sources which use them, this causes much better cache
matching when working with multiple builds.  For example testing code
with an without MALLOC_DEBUG will now use just over half the ccache
size, only main/astmm.o will have two builds cached instead of every

Reorder main/Makefile so _ASTCFLAGS set on specific object files are all
together, sorted by filename.  Stop adding -DMALLOC_DEBUG to CFLAGS of
bundled pjproject, this define is no longer used by any header so only
serves to break cache.

The only code change is a slight adjustment to how main/astmm.c is
initialized.  Initialization functions always exist so main/asterisk.c
can call them unconditionally.  Additionally rename the astmm
initialization functions so they are not exported.

Change-Id: Ie2085237a964f6e1e6fff55ed046e2afff83c027

18 months agoMerge "res_rtp_asterisk: In Developer Mode, do not require OpenSSL."
Joshua Colp [Wed, 1 Aug 2018 09:23:06 +0000 (04:23 -0500)]
Merge "res_rtp_asterisk: In Developer Mode, do not require OpenSSL."

18 months agoMerge "res_pjsip_pubsub: Use ast_true for "prune_on_boot"."
Joshua Colp [Tue, 31 Jul 2018 21:11:32 +0000 (16:11 -0500)]
Merge "res_pjsip_pubsub: Use ast_true for "prune_on_boot"."

18 months agopjsip_wizard.conf.sample: Update remote_hosts description.
Richard Mudgett [Tue, 31 Jul 2018 16:24:08 +0000 (11:24 -0500)]
pjsip_wizard.conf.sample: Update remote_hosts description.

Remove the note that SRV records are not supported as that is no longer


Change-Id: Id0dd6ef40e52702be9727a2b6122216cb00bb4ca

18 months agoCI: Add optional uninstall step before installing asterisk
George Joseph [Fri, 27 Jul 2018 18:23:02 +0000 (12:23 -0600)]
CI: Add optional uninstall step before installing asterisk

Change-Id: I7dedf1e925eafc3a0adf01dd9dfbe44eb642aab7

18 months agoBuildSystem: Enable ncurses for menuselect in Solaris 11.
Alexander Traud [Sat, 28 Jul 2018 16:49:17 +0000 (18:49 +0200)]
BuildSystem: Enable ncurses for menuselect in Solaris 11.

The check for the library ncurses should use not use the header <curses.h> but
<ncurses.h>, because on some platforms <curses.h> is not a drop-in replacement
for <ncurses.h>: For example in Solaris, the symbol initscr is a typedef in
<curses.h> to a symbol which does not exist in the library ncurses (initscr32).
Simply use <ncurses.h> when you link to ncurses.

Furthermore in Solaris, the header <ncurses.h> is in a subdirectory
/usr/include/ncurses and not available via pkg-config.


Change-Id: Ife367776b0ccf17d3fefed868245376bfb93745d

18 months agores_pjsip_pubsub: Use ast_true for "prune_on_boot".
Joshua Colp [Sat, 28 Jul 2018 13:00:40 +0000 (13:00 +0000)]
res_pjsip_pubsub: Use ast_true for "prune_on_boot".

Change-Id: Iedec4e7390b3e821987681da24d0298632b9873d

18 months agoBuildSystem: Enable Jansson in Solaris 11.
Alexander Traud [Sat, 28 Jul 2018 12:39:40 +0000 (14:39 +0200)]
BuildSystem: Enable Jansson in Solaris 11.

In Solaris, the header <jansson.h> is in /usr/include/jansson. To find
Jansson even in such a subdirectory, the tool pkg-config is queried via
AST_PKG_CONFIG_CHECK. For those platforms, which do not list Jansson via
pkg-config, the previous check remains and is executed thereafter.

Because the check for the NetBSD Editline library uses the tool pkg-config
the code of PKG_PROG_PKG_CONFIG must be used. Because that check happens
earlier than Jansson, it must be placed in front of that.


Change-Id: I69ea0f379f87a50049654b2487c76ee1c04fa53a

18 months agores_pjsip_endpoint_identifier_ip.c: Added regex support to match_header
Richard Mudgett [Tue, 24 Jul 2018 18:44:41 +0000 (13:44 -0500)]
res_pjsip_endpoint_identifier_ip.c: Added regex support to match_header

This patch adds regular expression support to make the identify section's
match_header option more useful when attempting to match complex headers
like the 'To' or 'From' headers.  The 'From' header has variable
components such as the tag parameter that you cannot predict.  To specify
a regular expression put slashes around the regular expression in place of
the header value.

match_header=From: /<sip:alice@127\\.0\\.0\\.1>/

* Added regex support to match_header so you could match a 'To' header
among other complex headers.

Fixed reported crashes when trying to match special headers like 'Contact'.
The identify section's match_header method used code that assumed you were
matching a generic header.  Any other type of header could cause a crash
if the header structure variant did not match the generic header enough.

* Made use code that will work for any header type instead of code
specific to generic headers.

Other fixes while in the area:

* Made check all headers of the requested name.
* Added some more sanity checks to the configured identify matching
options when applying the configuration.


Change-Id: I27dfd4ff5e2259b906640e3c330681b76b4ed1f1

18 months agores_pjsip_pubsub: Treat "prune_on_boot" as a yes / no.
Joshua Colp [Fri, 27 Jul 2018 15:46:12 +0000 (15:46 +0000)]
res_pjsip_pubsub: Treat "prune_on_boot" as a yes / no.

The alembic for the PJSIP subscription persistence table has the
"prune_on_boot" field as a boolean. While in Asterisk we are
tolerant of many different definitions of true and false in the
database we only accept "yes" and "no". This change makes the
field treated as a yes/no instead of an integer, thus storing
"yes" and "no" instead of "1" and "0".

Change-Id: Ic8b9211b36babefe78f70def6828a135a6ae7ab6

18 months agores_rtp_asterisk: In Developer Mode, do not require OpenSSL.
Alexander Traud [Fri, 27 Jul 2018 13:26:51 +0000 (15:26 +0200)]
res_rtp_asterisk: In Developer Mode, do not require OpenSSL.

OpenSSL is an optional external library and should stay optional even when
Developer Mode is configured.


Change-Id: Ia68a4cd5474b26d45e0f43b04032ad598022853b

18 months agoMerge "devicestate: Don't create topic when change isn't cached."
Joshua Colp [Fri, 27 Jul 2018 11:09:50 +0000 (06:09 -0500)]
Merge "devicestate: Don't create topic when change isn't cached."

18 months agoMerge "loader: Process dependencies for built-in modules."
Joshua Colp [Fri, 27 Jul 2018 11:04:51 +0000 (06:04 -0500)]
Merge "loader: Process dependencies for built-in modules."

18 months agoMerge "Build System: Create 'make install-configs' target."
Joshua Colp [Fri, 27 Jul 2018 10:51:44 +0000 (05:51 -0500)]
Merge "Build System: Create 'make install-configs' target."

18 months agoMerge "res_rtp_asterisk: Avoid merging command and regular T.140 text packets"
Joshua Colp [Fri, 27 Jul 2018 10:35:17 +0000 (05:35 -0500)]
Merge "res_rtp_asterisk: Avoid merging command and regular T.140 text packets"

18 months agoCI: Fix placement of job summary statments
George Joseph [Thu, 26 Jul 2018 23:54:36 +0000 (17:54 -0600)]
CI: Fix placement of job summary statments

Change-Id: Iace19e718f4e8fb48eb7dc9f98af53b115cc45f3

18 months agoloader: Process dependencies for built-in modules.
Corey Farrell [Thu, 26 Jul 2018 17:52:46 +0000 (13:52 -0400)]
loader: Process dependencies for built-in modules.

With the new module loader it was missed that built-in modules never
parsed dependencies from mod->info into vectors of mod.  This caused
manager to be initialized before acl (named_acl).  If manager.conf
used any named ACL's they would not be found and result in no ACL being
applied to the AMI user.

In addition to the manager ACL fix this adds "extconfig" to all builtin
modules which support realtime configuration.  This only matters if one
of the builtin modules is configured with 'preload', depending on
"extconfig" will cause config.c to automatically be initialize during
the preload stage.

Change-Id: I482ed6bca6c1064b05bb538d7861cd7a4f02d9fc

18 months agores_rtp_asterisk: Avoid merging command and regular T.140 text packets
neutrino88 [Wed, 18 Jul 2018 14:32:34 +0000 (10:32 -0400)]
res_rtp_asterisk: Avoid merging command and regular T.140 text packets

When realtime text packets are to be sent, the text is accumulated in a
buffer and sent regularly by a timer.  It can happen that commands such as
a backspace, CR, or LF get merged with regular text.  This breaks some

The proposed change:
* We test if the current packet contains a command.  If so we send the
buffer immediately.
* We test if the buffer contained a command.  If so we send the buffer
* We accumulate the text (or the command) in the buffer.


Change-Id: Ifbe993311410fa855cb8aa4a12084db75f413462

18 months agoMerge "CI: Add docker info to job summary"
Jenkins2 [Thu, 26 Jul 2018 16:44:30 +0000 (11:44 -0500)]
Merge "CI:  Add docker info to job summary"

18 months agoCI: Add docker info to job summary
George Joseph [Thu, 26 Jul 2018 16:34:20 +0000 (10:34 -0600)]
CI:  Add docker info to job summary

Change-Id: I45d52005a9b692ad303c11792f226ace1e449901

18 months agoBuild System: Create 'make install-configs' target.
Corey Farrell [Mon, 23 Jul 2018 18:49:33 +0000 (14:49 -0400)]
Build System: Create 'make install-configs' target.

This target requires specifying CONFIG_SRC=path_to_configs.  This can be
used to install custom configs for the Asterisk build while still
performing directory replacements on asterisk.conf.

Modify internal INSTALL_CONFIGS so first argument requires full path to
the config sources relative to Asterisk source root.

Change-Id: Idcd841df3c8d5bfe23d566bb9e2e448e9df4f8ab

18 months agojson.c: improve ast_json_to_ast_variables performance
Kevin Harwell [Wed, 25 Jul 2018 20:33:22 +0000 (15:33 -0500)]
json.c: improve ast_json_to_ast_variables performance

When converting from a json object to an ast variables list the conversion
algorithm was doing a complete traversal of the entire variables list for
every item appended from the json structure.

This patch makes it so the list is no longer traversed for each new ast
variable being appended.

Change-Id: I8bf496a1fc449485150d6db36bfc0354934a3977

18 months agodevicestate: Don't create topic when change isn't cached.
Joshua Colp [Wed, 25 Jul 2018 10:32:31 +0000 (10:32 +0000)]
devicestate: Don't create topic when change isn't cached.

When publishing a device state the change can be marked as being
cachable or not. If it is not cached the change is just published
to all interested and not stored away for later query. This was not
fully taken into account when publishing in stasis. The act of
publishing would create a topic for the device even if it may be

This change makes it so messages which are not cached won't create
a topic for the device. If a topic does already exist it will be
published to but otherwise the change will only be published to
the device state all topic.


Change-Id: I18da0e8cbb18e79602e731020c46ba4101e59f0a

18 months agoMerge "res_pjsip: Change log message from error to warning for valid use cases"
Joshua Colp [Wed, 25 Jul 2018 18:59:27 +0000 (13:59 -0500)]
Merge "res_pjsip: Change log message from error to warning for valid use cases"

18 months agoMerge "Build System: Silence build of bundled jansson."
George Joseph [Wed, 25 Jul 2018 18:59:22 +0000 (13:59 -0500)]
Merge "Build System: Silence build of bundled jansson."

18 months agoMerge "CI: Explicitly pass BRANCH_NAME to buildAsterisk and installAsterisk"
Jenkins2 [Wed, 25 Jul 2018 17:48:48 +0000 (12:48 -0500)]
Merge "CI: Explicitly pass BRANCH_NAME to buildAsterisk and installAsterisk"