3 years agoMerge changes from topic 'ASTERISK-27625'
George Joseph [Thu, 21 Jun 2018 15:26:31 +0000 (10:26 -0500)]
Merge changes from topic 'ASTERISK-27625'

* changes:
  channel.c: Make CHECK_BLOCKING() save thread LWP id for messages.
  channel.c: Fix usage of CHECK_BLOCKING()
  autoservice: Don't start channel autoservice if the thread is a user interface.

3 years agoMerge "ARI POST DTMF: Make not compete with channel's media thread."
George Joseph [Thu, 21 Jun 2018 15:26:23 +0000 (10:26 -0500)]
Merge "ARI POST DTMF: Make not compete with channel's media thread."

3 years agoMerge "AMI PlayDTMF Action: Make not compete with channel's media thread."
George Joseph [Thu, 21 Jun 2018 15:25:32 +0000 (10:25 -0500)]
Merge "AMI PlayDTMF Action: Make not compete with channel's media thread."

3 years agoMerge "menuselect/menuselect_curses: Resolves sprintf usage error"
Joshua Colp [Thu, 21 Jun 2018 09:24:35 +0000 (04:24 -0500)]
Merge "menuselect/menuselect_curses: Resolves sprintf usage error"

3 years agoMerge "Fix some doxygen and curly placement."
Jenkins2 [Wed, 20 Jun 2018 22:01:10 +0000 (17:01 -0500)]
Merge "Fix some doxygen and curly placement."

3 years agoMerge "Dialplan functions: Fix some channel autoservice misuse."
Jenkins2 [Wed, 20 Jun 2018 21:46:00 +0000 (16:46 -0500)]
Merge "Dialplan functions: Fix some channel autoservice misuse."

3 years agoMerge "tcptls.h: Remove redundant SSL_CTX typedef."
Jenkins2 [Wed, 20 Jun 2018 11:04:26 +0000 (06:04 -0500)]
Merge "tcptls.h: Remove redundant SSL_CTX typedef."

3 years agochannel.c: Make CHECK_BLOCKING() save thread LWP id for messages.
Richard Mudgett [Wed, 13 Jun 2018 16:33:44 +0000 (11:33 -0500)]
channel.c: Make CHECK_BLOCKING() save thread LWP id for messages.

* Removed an unnecessary call to ast_channel_blocker_set() in


Change-Id: I342168b999984666fb869cd519fe779583a73834

3 years agoARI POST DTMF: Make not compete with channel's media thread.
Richard Mudgett [Wed, 13 Jun 2018 21:41:43 +0000 (16:41 -0500)]
ARI POST DTMF: Make not compete with channel's media thread.

There can be one and only one thread handling a channel's media at a time.
Otherwise, we don't know which thread is going to handle the media frames.


Change-Id: I4d6a2fe7386ea447ee199003bf8ad681cb30454e

3 years agoAMI PlayDTMF Action: Make not compete with channel's media thread.
Richard Mudgett [Wed, 13 Jun 2018 18:05:03 +0000 (13:05 -0500)]
AMI PlayDTMF Action: Make not compete with channel's media thread.

There can be one and only one thread handling a channel's media at a time.
Otherwise, we don't know which thread is going to handle the media frames.


Change-Id: Ia341f1a6f4d54f2022261abec9021fe5b2eb4905

3 years agochannel.c: Fix usage of CHECK_BLOCKING()
Richard Mudgett [Tue, 12 Jun 2018 19:09:54 +0000 (14:09 -0500)]
channel.c: Fix usage of CHECK_BLOCKING()

The CHECK_BLOCKING() macro is used to indicate if a channel's handling
thread is about to do a blocking operation (poll, read, or write) of
media.  A few operations such as ast_queue_frame(), soft hangup, and
masquerades use the indication to wake up the blocked thread to reevaluate
what is going on.


Change-Id: I4dfc33e01e60627d962efa29d0a4244cf151a84d

3 years agoautoservice: Don't start channel autoservice if the thread is a user interface.
Richard Mudgett [Mon, 18 Jun 2018 23:04:54 +0000 (18:04 -0500)]
autoservice: Don't start channel autoservice if the thread is a user interface.

Executing dialplan functions from either AMI or ARI by getting a variable
could place the channel into autoservice.  However, these user interface
threads do not handle the channel's media so we wind up with two threads
attempting to handle the media.

There can be one and only one thread handling a channel's media at a time.
Otherwise, we don't know which thread is going to handle the media frames.


Change-Id: If2dc94ce15ddabf923ed1e2a65ea0ef56e013e49

3 years agoDialplan functions: Fix some channel autoservice misuse.
Richard Mudgett [Mon, 18 Jun 2018 21:07:47 +0000 (16:07 -0500)]
Dialplan functions: Fix some channel autoservice misuse.

* Fix off nominal paths leaving the channel in autoservice.
* Remove unnecessary start/stop channel autoservice.
* Fix channel locking around a channel datastore search.

Change-Id: I7ff2e42388064fe3149034ecae57604040b8b540

3 years agoFix some doxygen and curly placement.
Richard Mudgett [Tue, 19 Jun 2018 15:43:17 +0000 (10:43 -0500)]
Fix some doxygen and curly placement.

Change-Id: I9a784a7c804120a8fa826c2a4cb9957e4b0b2fc8

3 years agotcptls.h: Remove redundant SSL_CTX typedef.
Richard Mudgett [Mon, 18 Jun 2018 18:17:49 +0000 (13:17 -0500)]
tcptls.h: Remove redundant SSL_CTX typedef.

It is invalid to typedef something more than once.  Though not all gcc
compilers on different OS's complain about it.

Change-Id: I5a7d4565990c985822d61ce75bde0b45f9870540

3 years agochannel: Fix some more unprotected channel flag setting.
Richard Mudgett [Tue, 12 Jun 2018 20:13:14 +0000 (15:13 -0500)]
channel: Fix some more unprotected channel flag setting.

Change-Id: I34c3b1201b1de539945bcfdcb264fff30332d48c

3 years agoMerge "app_confbridge: Enable sending events to participants"
Jenkins2 [Mon, 18 Jun 2018 14:24:45 +0000 (09:24 -0500)]
Merge "app_confbridge:  Enable sending events to participants"

3 years agoMerge "app_mp3: remove 10 seconds of silence after mp3 playback"
Jenkins2 [Mon, 18 Jun 2018 12:43:09 +0000 (07:43 -0500)]
Merge "app_mp3: remove 10 seconds of silence after mp3 playback"

3 years agomenuselect/menuselect_curses: Resolves sprintf usage error
Matthew Fredrickson [Fri, 15 Jun 2018 20:21:27 +0000 (15:21 -0500)]
menuselect/menuselect_curses: Resolves sprintf usage error

Acccording to the man page for sprintf, using the same buffer for
output as one used as an input yields undefined behavior.
This patch should work around this problem.

Reported-by: Alexander Traud

Change-Id: I2213dcb454aff26457e2e4cc9c6821276463ae3a

3 years agoapp_mp3: remove 10 seconds of silence after mp3 playback
Sam Wierema [Tue, 12 Jun 2018 14:30:37 +0000 (16:30 +0200)]
app_mp3: remove 10 seconds of silence after mp3 playback

This patch changes the way asterisk polls output from mpg123, instead
of waiting for 10 seconds(when playing an http url) it now uses a
timeout of one second and iterates 10 times using this same timeout.

The main difference is that for every timeout asterisk receives it now
checks if mpg123 is still running before poll again.


Change-Id: Ib7df8462e3e380cb328011890ad9270d9e9b4620

3 years agoMerge "tests/test_utils: Repair ./configure --with-ssl=PATH."
Jenkins2 [Thu, 14 Jun 2018 16:32:45 +0000 (11:32 -0500)]
Merge "tests/test_utils: Repair ./configure --with-ssl=PATH."

3 years agoMerge "res_rtp_asterisk: Instead of ./configure use OPENSSL_NO_SRTP."
Joshua Colp [Thu, 14 Jun 2018 16:27:23 +0000 (11:27 -0500)]
Merge "res_rtp_asterisk: Instead of ./configure use OPENSSL_NO_SRTP."

3 years agotests/test_utils: Repair ./configure --with-ssl=PATH.
Alexander Traud [Wed, 13 Jun 2018 09:40:00 +0000 (11:40 +0200)]
tests/test_utils: Repair ./configure --with-ssl=PATH.


Change-Id: Ibcab8f556ee77776f203cff8b06d776a673b7bc4

3 years agochan_iax2: better handling for timeout and EINTR
ktyerman [Tue, 5 Jun 2018 01:31:39 +0000 (11:31 +1000)]
chan_iax2: better handling for timeout and EINTR

The iax2 module is not handling timeout and EINTR case properly. Mainly when
there is an interupt to the kernel thread. In case of ast_io_wait recieves a
signal, or timeout it can be an error or return 0 which eventually escapes the
thread loop, so that it cant recieve any data. This then causes the modules
receive queue to build up on the kernel and stop any communications via iax in

The proposed patch is for the iax module, so that timeout and EINTR does not
exit the thread.

Reported-by: Kirsty Tyerman

Change-Id: Ib4c32562f69335869adc1783608e940c3535fbfb

3 years agoapp_confbridge: Enable sending events to participants
George Joseph [Thu, 31 May 2018 21:22:13 +0000 (15:22 -0600)]
app_confbridge:  Enable sending events to participants

ConfBridge can now send events to participants via in-dialog MESSAGEs.
All current Confbridge events are supported, such as ConfbridgeJoin,
ConfbridgeLeave, etc.  In addition to those events, a new event
ConfbridgeWelcome has been added that will send a list of all
current participants to a new participant.

For all but the ConfbridgeWelcome event, the JSON message contains
information about the bridge, such as its id and name, and information
about the channel that triggered the event such as channel name,
callerid info, mute status, and the MSID labels for their audio and
video tracks. You can use the labels to correlate callerid and mute
status to specific video elements in a webrtc client.

To control this behavior, the following options have been added to

bridge_profile/enable_events:  This must be enabled on any bridge where
events are desired.

user_profile/send_events:  This must be set for a user profile to send
events.  Different user profiles connected to the same bridge can have
different settings.  This allows admins to get events but not normal
users for instance.

user_profile/echo_events:  In some cases, you might not want the user
triggering the event to get the event sent back to them.  To prevent it,
set this to false.

A change was also made to res_pjsip_sdp_rtp to save the generated msid
to the stream so it can be re-used.  This allows participant A's video
stream to appear as the same label to all other participants.

Change-Id: I26420aa9f101f0b2387dc9e2fd10733197f1318e

3 years agores_rtp_asterisk: Instead of ./configure use OPENSSL_NO_SRTP.
Alexander Traud [Wed, 13 Jun 2018 10:06:10 +0000 (12:06 +0200)]
res_rtp_asterisk: Instead of ./configure use OPENSSL_NO_SRTP.

Previously, Asterisk used its script ./configure, to test whether OpenSSL was
built with no-srtp (or was simply too old). However, the header file
<openssl/opensslconf.h> is the preferred way to detect the local configuration
of OpenSSL.

As a positive side-effect the script ./configure does not interleave the
detection of the Open Settlement Protocol Toolkit (OSPTK) with the detection of
individual features of OpenSSL anymore.

Change-Id: I3c77c7b00b2ffa2e935632097fa057b9fdf480c0

3 years agoMerge "rtp: Don't negotiate dynamic codecs using payload."
Jenkins2 [Tue, 12 Jun 2018 15:41:46 +0000 (10:41 -0500)]
Merge "rtp: Don't negotiate dynamic codecs using payload."

3 years agoMerge "res_rtp_asterisk: Allow OpenSSL configured with no-deprecated."
Jenkins2 [Tue, 12 Jun 2018 15:04:47 +0000 (10:04 -0500)]
Merge "res_rtp_asterisk: Allow OpenSSL configured with no-deprecated."

3 years agoMerge "crypto.h: Repair ./configure --with-ssl=PATH."
Jenkins2 [Tue, 12 Jun 2018 14:53:15 +0000 (09:53 -0500)]
Merge "crypto.h: Repair ./configure --with-ssl=PATH."

3 years agoMerge "res_crypto: Allow OpenSSL configured with no-deprecated."
Joshua Colp [Tue, 12 Jun 2018 13:41:30 +0000 (08:41 -0500)]
Merge "res_crypto: Allow OpenSSL configured with no-deprecated."

3 years agoMerge "res_srtp: Repair ./configure --with-ssl=PATH."
Jenkins2 [Tue, 12 Jun 2018 13:03:06 +0000 (08:03 -0500)]
Merge "res_srtp: Repair ./configure --with-ssl=PATH."

3 years agoMerge "func_odbc: NODATA if SQLNumResultCols returned 0 columns on readsql"
Jenkins2 [Tue, 12 Jun 2018 12:51:26 +0000 (07:51 -0500)]
Merge "func_odbc: NODATA if SQLNumResultCols returned 0 columns on readsql"

3 years agortp: Don't negotiate dynamic codecs using payload.
Joshua Colp [Tue, 5 Jun 2018 09:36:35 +0000 (09:36 +0000)]
rtp: Don't negotiate dynamic codecs using payload.

In Asterisk there are some dynamic codecs that have
a fixed payload number. This number was being improperly
used to negotiate the codec, instead of using the name
and sample rate. This could result in the wrong payload
number being negotiated for a codec.

This change makes it so that only static payloads
will be negotiated using their payload number.


Change-Id: Ia865830170fd3f808cdb33104f3d4c4ffdc77570

3 years agoMerge "chan_pjsip: Register for "BEFORE_MEDIA" responses"
Jenkins2 [Mon, 11 Jun 2018 23:14:14 +0000 (18:14 -0500)]
Merge "chan_pjsip:  Register for "BEFORE_MEDIA" responses"

3 years agoMerge "AST-2018-008: Fix enumeration of endpoints from ACL rejected addresses."
Kevin Harwell [Mon, 11 Jun 2018 19:34:36 +0000 (14:34 -0500)]
Merge "AST-2018-008: Fix enumeration of endpoints from ACL rejected addresses."

3 years agoAST-2018-007: iostreams potential DoS when client connection closed prematurely
Sean Bright [Mon, 16 Apr 2018 19:13:58 +0000 (15:13 -0400)]
AST-2018-007: iostreams potential DoS when client connection closed prematurely

Before Asterisk sends an HTTP response (at least in the case of errors),
it attempts to read & discard the content of the request. If the client
lies about the Content-Length, or the connection is closed from the
client side before "Content-Length" bytes are sent, the request handling
thread will busy loop.


Change-Id: I945c5fc888ed92be625b8c35039fc6d2aa89c762

3 years agoAST-2018-008: Fix enumeration of endpoints from ACL rejected addresses.
Richard Mudgett [Mon, 30 Apr 2018 22:38:58 +0000 (17:38 -0500)]
AST-2018-008: Fix enumeration of endpoints from ACL rejected addresses.

When endpoint specific ACL rules block a SIP request they respond with a
403 forbidden.  However, if an endpoint is not identified then a 401
unauthorized response is sent.  This vulnerability just discloses which
requests hit a defined endpoint.  The ACL rules cannot be bypassed to gain
access to the disclosed endpoints.

* Made endpoint specific ACL rules now respond with a 401 unauthorized
which is the same as if an endpoint were not identified.  The fix is
accomplished by replacing the found endpoint with the artificial endpoint
which always fails authentication.


Change-Id: Icb275a54ff8e2df6c671a6d9bda37b5d732b3b32

3 years agores_rtp_asterisk: Allow OpenSSL configured with no-deprecated.
Alexander Traud [Fri, 8 Jun 2018 20:02:38 +0000 (22:02 +0200)]
res_rtp_asterisk: Allow OpenSSL configured with no-deprecated.

Furthermore, allow OpenSSL configured with no-dh. Additionally, this change
allows auto-negotiation of the elliptic curve/group for servers, not only with
OpenSSL 1.0.2 but also with OpenSSL 1.1.0 and newer. This enables X25519
(since OpenSSL 1.1.0) and X448 (since OpenSSL 1.1.1) as a side-effect.


Change-Id: I5b0dd47c5194ee17f830f869d629d7ef212cf537

3 years agocrypto.h: Repair ./configure --with-ssl=PATH.
Alexander Traud [Fri, 8 Jun 2018 11:01:53 +0000 (13:01 +0200)]
crypto.h: Repair ./configure --with-ssl=PATH.


Change-Id: Iac49d9f82faeb8a4611c6805906bd6d650b1b1d8

3 years agores_crypto: Allow OpenSSL configured with no-deprecated.
Alexander Traud [Fri, 8 Jun 2018 09:03:35 +0000 (11:03 +0200)]
res_crypto: Allow OpenSSL configured with no-deprecated.

The header <openssl/rsa.h> had to be included explicitly.


Change-Id: I41743801eed998c039d73db7a0762d104a4f75b2

3 years agores_srtp: Repair ./configure --with-ssl=PATH.
Alexander Traud [Fri, 8 Jun 2018 07:41:01 +0000 (09:41 +0200)]
res_srtp: Repair ./configure --with-ssl=PATH.


Change-Id: Ibb7dc148a0048f4f9c3b12937ba4240dff0d15e2

3 years agofunc_odbc: NODATA if SQLNumResultCols returned 0 columns on readsql
Alexei Gradinari [Thu, 31 May 2018 15:25:40 +0000 (11:25 -0400)]
func_odbc: NODATA if SQLNumResultCols returned 0 columns on readsql

The functions acf_odbc_read/cli_odbc_read ignore a number of columns
returned by the SQLNumResultCols.
If the number of columns is zero it means no data.
In this case, a SQLFetch function has to be not called,
because it will cause an error.

ASTERISK-27888 #close

Change-Id: Ie0f7bdac6c405aa5bbd38932c7b831f90729ee19

3 years agochan_pjsip: Register for "BEFORE_MEDIA" responses
George Joseph [Thu, 7 Jun 2018 13:46:03 +0000 (07:46 -0600)]
chan_pjsip:  Register for "BEFORE_MEDIA" responses

chan_pjsip wasn't registering for "BEFORE_MEDIA" responses which meant
it was not updating HANGUPCAUSE for 4XX responses.  If the remote end
sent a "180 Ringing", then a "486 Busy", the hangup cause was left at
"180 Normal Clearing".

* Removed chan_pjsip_incoming_response from the original session
  supplement (which was handling only "AFTER MEDIA") and added it to a
  new session supplement which accepts both "BEFORE_MEDIA" and

* Also cleaned up some cleanup code in load module.


Change-Id: If9b860541887aca8ac2c9f2ed51ceb0550fb007a

3 years agoooh323c: GCC 8.1 warned about output truncated before terminating nul.
Alexander Traud [Thu, 7 Jun 2018 12:19:39 +0000 (14:19 +0200)]
ooh323c: GCC 8.1 warned about output truncated before terminating nul.


Change-Id: I5a8e894f4924ef52e3094f6870656a559d67f3d7

3 years agoMerge "pjsip_options: handle modification of qualify options in realtime"
George Joseph [Wed, 6 Jun 2018 15:12:58 +0000 (10:12 -0500)]
Merge "pjsip_options: handle modification of qualify options in realtime"

3 years agoMerge "pjsip_options: show/reload AOR qualify options using CLI"
George Joseph [Wed, 6 Jun 2018 15:11:06 +0000 (10:11 -0500)]
Merge "pjsip_options: show/reload AOR qualify options using CLI"

3 years agoMerge "app_confbridge: Add talking indicator for ConfBridgeList AMI response"
George Joseph [Wed, 6 Jun 2018 14:47:19 +0000 (09:47 -0500)]
Merge "app_confbridge: Add talking indicator for ConfBridgeList AMI response"

3 years agoMerge "cdr_mysql: my_connect_db(): reduce indentation"
George Joseph [Wed, 6 Jun 2018 13:12:48 +0000 (08:12 -0500)]
Merge "cdr_mysql: my_connect_db(): reduce indentation"

3 years agoMerge "cdr_mysql: split mysql init out of my_load_module"
George Joseph [Wed, 6 Jun 2018 13:12:38 +0000 (08:12 -0500)]
Merge "cdr_mysql: split mysql init out of my_load_module"

3 years agoMerge "bridge_channel.c: Fix Deadlock when using Local channels and fax gateway"
Joshua Colp [Wed, 6 Jun 2018 10:46:47 +0000 (05:46 -0500)]
Merge "bridge_channel.c: Fix Deadlock when using Local channels and fax gateway"

3 years agoMerge "tcptls: Allow OpenSSL configured with no-dh."
Joshua Colp [Wed, 6 Jun 2018 09:36:06 +0000 (04:36 -0500)]
Merge "tcptls: Allow OpenSSL configured with no-dh."

3 years agoMerge "tcptls.h: Repair ./configure --with-ssl=PATH."
George Joseph [Tue, 5 Jun 2018 19:21:15 +0000 (14:21 -0500)]
Merge "tcptls.h: Repair ./configure --with-ssl=PATH."

3 years agopjsip_options: show/reload AOR qualify options using CLI
Alexei Gradinari [Tue, 5 Jun 2018 18:43:18 +0000 (14:43 -0400)]
pjsip_options: show/reload AOR qualify options using CLI

Currentrly pjsip_options code does not handle the situation when the
AOR qualify options were changed.

Also there is no way to find out what qualify options are using.

This patch add CLI commands to show and synchronize Aor qualify options:
pjsip show qualify endpoint <id>
    Show the current qualify options for all Aors on the PJSIP endpoint.
pjsip show qualify aor <id>
    Show the PJSIP Aor current qualify options.
pjsip reload qualify endpoint <id>
    Synchronize the qualify options for all Aors on the PJSIP endpoint.
pjsip reload qualify aor <id>
    Synchronize the PJSIP Aor qualify options.


Change-Id: I1746d10ef2b7954f2293f2e606cdd7428068c38c

3 years agopjsip_options: handle modification of qualify options in realtime
Alexei Gradinari [Tue, 22 May 2018 21:21:10 +0000 (17:21 -0400)]
pjsip_options: handle modification of qualify options in realtime

Currentrly pjsip_options code does not handle the situation when the
qualify options were changed in realtime database.
Only 'module reload res_pjsip' helps.

This patch add a check on contact add/update observers if the contact
qualify options are different than local aor qualify options.
If the qualify options were modified then synchronize
the pjsip_options AOR local state.


Change-Id: Id55210a18e62ed5d35a88e408d5fe84a3c513c62

3 years agoMerge "tcptls: Allow OpenSSL 1.1.x configured with enable-ssl3-method no-deprecated."
George Joseph [Tue, 5 Jun 2018 18:01:31 +0000 (13:01 -0500)]
Merge "tcptls: Allow OpenSSL 1.1.x configured with enable-ssl3-method no-deprecated."

3 years agoMerge "app_meetme: Fix manager event documentation for several events."
Joshua Colp [Tue, 5 Jun 2018 11:54:00 +0000 (06:54 -0500)]
Merge "app_meetme: Fix manager event documentation for several events."

3 years agobridge_channel.c: Fix Deadlock when using Local channels and fax gateway
Pirmin Walthert [Wed, 30 May 2018 06:12:30 +0000 (08:12 +0200)]
bridge_channel.c: Fix Deadlock when using Local channels and fax gateway

ast_indicate is invoked with the bridge locked. As ast_indicate locks the
other end of the bridge as well this can lead to a deadlock in some situations.
(Especially when a different thread does the same in the reverse order).
This patch calls ast_indicate after unlocking the bridge which fixes the
deadlock. Calling ast_indicate with these parameters without locking the
bridge should be safe as this is done at different places without a
bridge lock.

ASTERISK-27094 #close
Reported-by: David Brillert

Change-Id: I5f86c1e2ce75b9929a36ab589b18c450e62ea35f

3 years agoapp_sendtext: Allow content types other than text/plain
George Joseph [Mon, 4 Jun 2018 14:50:51 +0000 (08:50 -0600)]
app_sendtext:  Allow content types other than text/plain

There was no real reason to limit the conteny type to text/plain other
than that's what it was limited to before.  Now any text/* content
type will be allowed for channel drivers that don't support enhanced
messaging and any type will be allowed for channel drivers that do
support enhanced messaging.

Change-Id: I94a90cfee98b4bc8e22aa5c0b6afb7b862f979d9

3 years agoapp_confbridge: Add talking indicator for ConfBridgeList AMI response
William McCall [Tue, 29 May 2018 00:17:52 +0000 (00:17 +0000)]
app_confbridge: Add talking indicator for ConfBridgeList AMI response

When an AMI client connects, it cannot determine if a user was talking
prior to a transition in the user speaking state (which would generate
a ConfbridgeTalking event). This patch causes app_confbridge to track the
talking state and make this state available via ConfBridgeList.

ASTERISK-27877 #close

Change-Id: I19b5284f34966c3fda94f5b99a7e40e6b89767c6

3 years agoMerge "ast_coredumper: Fix output directory and variable precedence"
Joshua Colp [Thu, 31 May 2018 10:16:08 +0000 (05:16 -0500)]
Merge "ast_coredumper:  Fix output directory and variable precedence"

3 years agoapp_meetme: Fix manager event documentation for several events.
Richard Mudgett [Tue, 29 May 2018 17:28:48 +0000 (12:28 -0500)]
app_meetme: Fix manager event documentation for several events.

The MeetmeJoin, MeetmeLeave, MeetmeEnd, MeetmeMute, MeetmeTalking, and
MeetmeTalkRequest AMI events were documented with sending out a Usernum
header when the User header was actually output.

* Change the online documentation to match reality.


Change-Id: I437bc70618d07c183c9624b7069c2fcae7f17a39

3 years agoMerge "libasteriskssl: Allow OpenSSL 1.0.2 configured with no-deprecated."
Joshua Colp [Tue, 29 May 2018 17:07:51 +0000 (12:07 -0500)]
Merge "libasteriskssl: Allow OpenSSL 1.0.2 configured with no-deprecated."

3 years agotcptls.h: Repair ./configure --with-ssl=PATH.
Alexander Traud [Mon, 28 May 2018 15:29:23 +0000 (17:29 +0200)]
tcptls.h: Repair ./configure --with-ssl=PATH.

asterisk/tcptls.h was included (explicitly, implicitly, or transitively). Those
inclusions got replaced by forward declarations. As side effect, the inclusions
got completed.


Change-Id: I9d102728e30336d6522e5e4ae9e964013a0835f7

3 years agotcptls: Allow OpenSSL configured with no-dh.
Alexander Traud [Fri, 25 May 2018 14:55:26 +0000 (16:55 +0200)]
tcptls: Allow OpenSSL configured with no-dh.

Additionally, this change allows auto-negotiation of the elliptic curve/group
for servers, not only with OpenSSL 1.0.2 but also with OpenSSL 1.1.0 and newer.
This enables X25519 (since OpenSSL 1.1.0) and X448 (since OpenSSL 1.1.1) as a


Change-Id: I62c2aba4a630aefc231b71f646207e8c027d9497

3 years agotcptls: Allow OpenSSL 1.1.x configured with enable-ssl3-method no-deprecated.
Alexander Traud [Fri, 25 May 2018 12:22:14 +0000 (14:22 +0200)]
tcptls: Allow OpenSSL 1.1.x configured with enable-ssl3-method no-deprecated.


Change-Id: Ica65113511c7a1c13f7988e7d9e7d9e7f3f620dd

3 years agoMerge "rtp: Add support for RTP extension negotiation and abs-send-time."
Joshua Colp [Thu, 24 May 2018 20:26:57 +0000 (15:26 -0500)]
Merge "rtp: Add support for RTP extension negotiation and abs-send-time."

3 years agoMerge "res/res_rtp_asterisk: ensure marker bit is correctly set on ssrc change"
Joshua Colp [Thu, 24 May 2018 20:09:58 +0000 (15:09 -0500)]
Merge "res/res_rtp_asterisk: ensure marker bit is correctly set on ssrc change"

3 years agoast_coredumper: Fix output directory and variable precedence
George Joseph [Tue, 15 May 2018 13:45:20 +0000 (07:45 -0600)]
ast_coredumper:  Fix output directory and variable precedence

The OUTPUTDIR variable in ast_debug_tools.conf.sample is now set
to "/tmp" instead of "/some/directory".

Variables set on the command line or that are already in the
environment now take predecence over variables set in the config files.

Reported by: Ted G

Change-Id: Ie8baec52d531886bf5849ec1d59bb59dc87ad387

3 years agoMerge "tcptls: Repair ./configure --with-ssl=PATH."
Joshua Colp [Thu, 24 May 2018 11:20:15 +0000 (06:20 -0500)]
Merge "tcptls: Repair ./configure --with-ssl=PATH."

3 years agoMerge "app_queue: Update year Copyright and fix missing tabs in documentation"
Joshua Colp [Thu, 24 May 2018 10:49:41 +0000 (05:49 -0500)]
Merge "app_queue: Update year Copyright and fix missing tabs in documentation"

3 years agoMerge "config.c: Fix successful DELETE treated as failure"
Joshua Colp [Thu, 24 May 2018 10:49:21 +0000 (05:49 -0500)]
Merge "config.c: Fix successful DELETE treated as failure"

3 years agoMerge "channel.c: Fix off nominal channel allocation failure path."
Joshua Colp [Thu, 24 May 2018 10:18:16 +0000 (05:18 -0500)]
Merge "channel.c: Fix off nominal channel allocation failure path."

3 years agores/res_rtp_asterisk: ensure marker bit is correctly set on ssrc change
Torrey Searle [Wed, 9 May 2018 13:31:47 +0000 (15:31 +0200)]
res/res_rtp_asterisk: ensure marker bit is correctly set on ssrc change

Certain race conditions between changing bridge types and DTMF can
cause the current FLAG_NEED_MARKER_BIT to send the marker bit before
the actual first packet of native bridging.

This logic keeps track of the ssrc the bridge is currently sending
and will correctly ensure the marker bit is set if SSRC as changed
from the previous sent packet.


Change-Id: I01858bd0235f1e5e629e20de71b422b16f55759b

3 years agoMerge "netsock2: Add ast_sockaddr_resolve_first_af to netsock2 public API"
Joshua Colp [Wed, 23 May 2018 17:10:13 +0000 (12:10 -0500)]
Merge "netsock2: Add ast_sockaddr_resolve_first_af to netsock2 public API"

3 years agortp: Add support for RTP extension negotiation and abs-send-time.
Joshua Colp [Mon, 23 Apr 2018 14:04:01 +0000 (14:04 +0000)]
rtp: Add support for RTP extension negotiation and abs-send-time.

When RTP was originally created it had the ability to place a single
extension in an RTP packet. In practice people wanted to potentially
put multiple extensions in one and so RFC 5285 (obsoleted by RFC
8285) came into existence. This allows RTP extensions to be negotiated
with a unique identifier to be used in the RTP packet, allowing
multiple extensions to be present in the packet.

This change extends the RTP engine API to add support for this. A
user of it can enable extensions and the API provides the ability to
retrieve the information (to construct SDP for example) and to provide
negotiated information (from SDP). The end result is that the RTP
engine can then query to see if the extension has been negotiated and
what unique identifier is to be used. It is then up to the RTP engine
implementation to construct the packet appropriately.

The first extension to use this support is abs-send-time which is
defined in the REMB draft[1] and is a second timestamp placed in an
RTP packet which is for when the packet has left the sending system.
It is used to more accurately determine the available bandwidth.



Change-Id: I508deac557867b1e27fc7339be890c8018171588

3 years agochannel.c: Fix off nominal channel allocation failure path.
Richard Mudgett [Tue, 22 May 2018 22:17:31 +0000 (17:17 -0500)]
channel.c: Fix off nominal channel allocation failure path.

__ast_channel_alloc_ap() had a failure exit path that hadn't setup the fd
descriptors to -1 yet.  The destructor would then attempt to close these
fd's that had never been opened.

Change-Id: Icf21093f36c60781e8cf6ee9d586536302af33e3

3 years agoapp_queue: Update year Copyright and fix missing tabs in documentation
Rodrigo Ramírez Norambuena [Fri, 18 May 2018 17:46:03 +0000 (13:46 -0400)]
app_queue: Update year Copyright and fix missing tabs in documentation

Change-Id: Ieb8faf37dc765463ee5dbca1d1343242c756b1c7

3 years agoconfig.c: Fix successful DELETE treated as failure
Alexei Gradinari [Fri, 18 May 2018 21:45:22 +0000 (17:45 -0400)]
config.c: Fix successful DELETE treated as failure

The config engine destroy_func callback function returns the number of
rows deleted or -1 on error.  But the function
ast_destroy_realtime_fields treated non-zero return values as error.


Change-Id: Ied02b38e8196cb03043e609a0679feebd288d17b

3 years agonetsock2: Add ast_sockaddr_resolve_first_af to netsock2 public API
Matthew Fredrickson [Mon, 14 May 2018 11:07:34 +0000 (06:07 -0500)]
netsock2: Add ast_sockaddr_resolve_first_af to netsock2 public API

This function originally was used in chan_sip to enable some simplifying
assumptions and eventually was copy and pasted into res_pjsip_logger and
res_hep.  Since it's replicated in three places, it's probably best to
move it into the public netsock2 API for these modules to use.

Change-Id: Id52e23be885601c51d70259f62de1a5e59d38d04

3 years agoMerge "app_voicemail: Fix data-type mismatch between app_voicemail and database"
Joshua Colp [Mon, 21 May 2018 14:05:19 +0000 (09:05 -0500)]
Merge "app_voicemail: Fix data-type mismatch between app_voicemail and database"

3 years agolibasteriskssl: Allow OpenSSL 1.0.2 configured with no-deprecated.
Alexander Traud [Sun, 20 May 2018 11:36:51 +0000 (13:36 +0200)]
libasteriskssl: Allow OpenSSL 1.0.2 configured with no-deprecated.

Use CRYPTO_set_id_callback(.) only with OpenSSL 0.9.8 and older.


Change-Id: Iadd58d5bf6f538eb224203970a4e88e26f259655

3 years agotcptls: Repair ./configure --with-ssl=PATH.
Alexander Traud [Sat, 19 May 2018 13:23:30 +0000 (15:23 +0200)]
tcptls: Repair ./configure --with-ssl=PATH.

SSL_OP_NO_TLSv1_1 and SSL_OP_NO_TLSv1_2 got discovered without honoring a PATH.


Change-Id: I8cd358eed7411726d08fa7b01691bef122fbeb71

3 years agoMerge "chan_mobile: support handling of caller-id names ("cnam")."
Kevin Harwell [Fri, 18 May 2018 22:44:23 +0000 (17:44 -0500)]
Merge "chan_mobile: support handling of caller-id names ("cnam")."

3 years agoMerge "app_voicemail: Fix incorrect msg leaving/retrieving an ODBC voicemail"
Kevin Harwell [Fri, 18 May 2018 22:18:59 +0000 (17:18 -0500)]
Merge "app_voicemail: Fix incorrect msg leaving/retrieving an ODBC voicemail"

3 years agoMerge "rtp_engine: Allow Media Formats with add_static_payload(-1) on egress again."
Kevin Harwell [Fri, 18 May 2018 21:42:29 +0000 (16:42 -0500)]
Merge "rtp_engine: Allow Media Formats with add_static_payload(-1) on egress again."

3 years agoapp_voicemail: Fix data-type mismatch between app_voicemail and database
Nic Colledge [Tue, 27 Mar 2018 23:53:07 +0000 (00:53 +0100)]
app_voicemail: Fix data-type mismatch between app_voicemail and database

Fix data-type mismatch between app_voicemail and database columns
exposed by new version of MariaDB


Change-Id: I8543ad480a08c98be78bde1ee870e6e6c84b2c5b

3 years agoapp_voicemail: Fix incorrect msg leaving/retrieving an ODBC voicemail
Nic Colledge [Sat, 12 May 2018 11:53:13 +0000 (12:53 +0100)]
app_voicemail: Fix incorrect msg leaving/retrieving an ODBC voicemail

Correct the log warning message shown when ODBC voicemail
retrieve_file is called and there is a null value in the category
A more meaningfull message is now written at debug level.


Change-Id: Ic36e97d5eb070a23a12ba45972f6b53e2182a3f4

3 years agochan_mobile: support handling of caller-id names ("cnam").
Brian P. Martin [Wed, 18 Apr 2018 02:15:08 +0000 (19:15 -0700)]
chan_mobile: support handling of caller-id names ("cnam").

Add support to handle caller-ID names ("cnam") in addition to caller-ID
numbers.  The prior code ignored the caller-ID name altogether, and
used the local name for the cell phone (e.g. "my-iphone") in its place.

Note: as of this writing, at least some Android phones don't pass cnam to
us. This can be seen by issuing "core set debug 2" in the CLI and watching
the "CLIP" record when a call comes in.  If cnam isn't in the CLIP record,
there's nothing we can do to provide one.  We'll provide a null cnam field,
so later Asterisk processes know to try other sources (e.g. cidname database,
OpenCNAM, etc.).

Reported by: Brian Martin
Tested by: Brian Martin

Change-Id: I89490d85fa406c36261879c50ae5e65595538ba5

3 years agores_pjsip_endpoint_identifier_ip: Unregister the module for headers.
Alexander Traud [Thu, 17 May 2018 06:58:43 +0000 (08:58 +0200)]
res_pjsip_endpoint_identifier_ip: Unregister the module for headers.

Asterisk uses Reference Counting to track whether a module can be unloaded.
Every consumer who requires a module, increases the reference count. When the
consumer goes, is unloaded itself, it has to decrease the reference count on
all its used/required modules. That way
 core stop gracefully
works on the command-line interface (CLI): One module after the other is
unloaded. A recent change broke this for the module res_pjsip.


Change-Id: I261abcb411d026bbb0691cc78f28300bfd3103a3

3 years agoMerge "cli: Display correct unit for HTTP timeout in "manager show settings"."
Joshua Colp [Wed, 16 May 2018 18:56:48 +0000 (13:56 -0500)]
Merge "cli: Display correct unit for HTTP timeout in "manager show settings"."

3 years agoMerge "Fix GCC 8 build issues."
Joshua Colp [Wed, 16 May 2018 18:56:34 +0000 (13:56 -0500)]
Merge "Fix GCC 8 build issues."

3 years agoMerge "rtp_engine: Remove the double assigned RTP payload ID of H.263+."
Joshua Colp [Tue, 15 May 2018 09:06:39 +0000 (04:06 -0500)]
Merge "rtp_engine: Remove the double assigned RTP payload ID of H.263+."

3 years agoMerge "git: Ignore *.orig."
Joshua Colp [Mon, 14 May 2018 11:56:08 +0000 (06:56 -0500)]
Merge "git: Ignore *.orig."

3 years agoMerge "rtp_engine: Avoid a typo error in Doxygen for ast_rtp_codecs_find_payload_code."
Joshua Colp [Mon, 14 May 2018 11:30:44 +0000 (06:30 -0500)]
Merge "rtp_engine: Avoid a typo error in Doxygen for ast_rtp_codecs_find_payload_code."

3 years agoMerge "pjsip: Rewrite OPTIONS support with new eyes."
Joshua Colp [Mon, 14 May 2018 09:06:53 +0000 (04:06 -0500)]
Merge "pjsip: Rewrite OPTIONS support with new eyes."

3 years agortp_engine: Remove the double assigned RTP payload ID of H.263+.
Alexander Traud [Fri, 11 May 2018 17:49:12 +0000 (19:49 +0200)]
rtp_engine: Remove the double assigned RTP payload ID of H.263+.

Mantis-3709 (Commit 68ff3c3, Asterisk 1.2) added support for the video format
H.263+. For this, the RTP payload ID 103 got assigned statically. Commit f1aadc8
assigned another payload ID 98 for this format in Asterisk 1.6.

Change-Id: I90e35b158487f8f1f8187da6241b54cd3b74e667

3 years agocli: Display correct unit for HTTP timeout in "manager show settings".
Corey Farrell [Fri, 11 May 2018 17:26:39 +0000 (13:26 -0400)]
cli: Display correct unit for HTTP timeout in "manager show settings".

HTTP timeout is in seconds, not minutes.

ASTERISK-27852 #close

Change-Id: Ie6640835cb07307555741f9b559c2eb876d9343e

3 years agortp_engine: Avoid a typo error in Doxygen for ast_rtp_codecs_find_payload_code.
Alexander Traud [Fri, 11 May 2018 15:37:57 +0000 (17:37 +0200)]
rtp_engine: Avoid a typo error in Doxygen for ast_rtp_codecs_find_payload_code.

Change-Id: Ica089d4507a27ddfc4ce3a88d697ffbef378de48

3 years agoFix GCC 8 build issues.
Corey Farrell [Mon, 7 May 2018 02:17:34 +0000 (22:17 -0400)]
Fix GCC 8 build issues.

This fixes build warnings found by GCC 8.  In some cases format
truncation is intentional so the warning is just suppressed.

ASTERISK-27824 #close

Change-Id: I724f146cbddba8b86619d4c4a9931ee877995c84

3 years agortp_engine: Allow Media Formats with add_static_payload(-1) on egress again.
Alexander Traud [Fri, 11 May 2018 12:10:51 +0000 (14:10 +0200)]
rtp_engine: Allow Media Formats with add_static_payload(-1) on egress again.

This issue affected only installations with rtp_use_dynamic=yes in asterisk.conf
which is the default since Asterisk 15. Codec 2 and SiLK were built-in examples
of media formats which were affected.

Reported by: Dinis Brazão, Selene Feigl

Change-Id: I08c1e76433a67e4350141d38cacf3a1cb5086496