asterisk/asterisk.git
4 years agores_pjsip: Add an 'auto' option for DTMF Mode
Matthew Jordan [Fri, 10 Apr 2015 17:56:47 +0000 (17:56 +0000)]
res_pjsip: Add an 'auto' option for DTMF Mode

This patch adds support for automatically detecting the type of DTMF that a
PJSIP endpoint supports. When the 'dtmf_mode' endpoint option is set to 'auto',
the channel created for an endpoint will attempt to determine if RFC 4733
DTMF is supported. If so, it will use that DTMF type. If not, the DTMF type
for the channel will be set to inband.

Review: https://reviewboard.asterisk.org/r/4438

ASTERISK-24706 #close
Reported by: yaron nahum
patches:
  yaron_patch_3_Feb.diff submitted by yaron nahum (License 6676)
........

Merged revisions 434637 from http://svn.asterisk.org/svn/asterisk/branches/13

git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@434638 65c4cc65-6c06-0410-ace0-fbb531ad65f3

4 years agores_pjsip_config_wizard: Cleanup load unload
George Joseph [Fri, 10 Apr 2015 17:00:38 +0000 (17:00 +0000)]
res_pjsip_config_wizard: Cleanup load unload

While investigating other unload issues I realized that the load/unload process
for the config wizard was pretty ugly so I've refactored it as follows...

When the res_pjsip sorcery instance is created the config_wizard bumps it's own
module reference to prevent it from unloading while the sorcery instance is
still active.  When res_pjsip unloads and it's sorcery instance is destroyed,
the config wizard unrefs itself which then allows itself to unload cleanly.
Since the config wizard now can't load after res_pjsip or unload before it
(which should have been the correct behavior all along), I was able to remove
the chunks of code in both load_module and unload_module that handled that case.

Ran the testsuite tests to insure there were no functional changes and REF_DEBUG
to insure that Asterisk was shutting down cleanly with no FRACKs or leaks.

Tested-by: George Joseph
Review: https://reviewboard.asterisk.org/r/4610/
........

Merged revisions 434619 from http://svn.asterisk.org/svn/asterisk/branches/13

git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@434620 65c4cc65-6c06-0410-ace0-fbb531ad65f3

4 years agobridge_softmix.c,channel.c: Minor code simplification and cleanup.
Richard Mudgett [Fri, 10 Apr 2015 16:38:58 +0000 (16:38 +0000)]
bridge_softmix.c,channel.c: Minor code simplification and cleanup.

* Made code easier to follow in bridge_softmix.c:analyse_softmix_stats()
and made some debug messages more helpful.

* Made some debug and warning messages more helpful in
channel.c:set_format().

Review: https://reviewboard.asterisk.org/r/4607/
........

Merged revisions 434617 from http://svn.asterisk.org/svn/asterisk/branches/13

git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@434618 65c4cc65-6c06-0410-ace0-fbb531ad65f3

4 years agotranslate.c: Only select audio codecs to determine the best translation choice.
Richard Mudgett [Fri, 10 Apr 2015 16:32:28 +0000 (16:32 +0000)]
translate.c: Only select audio codecs to determine the best translation choice.

Given a source capability of h264 and ulaw, a destination capability of
h264 and g722 then ast_translator_best_choice() would pick h264 as the
best choice even though h264 is a video codec and Asterisk only supports
translation of audio codecs.  When the audio starts flowing, there are
warnings about a codec mismatch when the channel tries to write a frame to
the peer.

* Made ast_translator_best_choice() only select audio codecs.

* Restore a check in channel.c:set_format() lost after v1.8 to prevent
trying to set a non-audio codec.

This is an intermediate patch for a series of patches aimed at improving
translation path choices for ASTERISK-24841.

This patch is a complete enough fix for ASTERISK-21777 as the v11 version
of ast_translator_best_choice() does the same thing.  However, chan_sip.c
still somehow tries to call ast_codec_choose() which then calls
ast_best_codec() with a capability set that doesn't contain any audio
formats for the incoming call.  The remaining warning message seems to be
a benign transient.

ASTERISK-21777 #close
Reported by: Nick Ruggles

ASTERISK-24380 #close
Reported by: Matt Jordan

Review: https://reviewboard.asterisk.org/r/4605/
........

Merged revisions 434614 from http://svn.asterisk.org/svn/asterisk/branches/11
........

Merged revisions 434615 from http://svn.asterisk.org/svn/asterisk/branches/13

git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@434616 65c4cc65-6c06-0410-ace0-fbb531ad65f3

4 years agores/ari: Fix model validation for ChannelHold event
Matthew Jordan [Fri, 10 Apr 2015 14:56:05 +0000 (14:56 +0000)]
res/ari: Fix model validation for ChannelHold event

When the ChannelHold event was added, the 'musicclass' parameter was
erroneously removed. This caused the ChannelHold events to be rejected as
they failed model validation. This patch updates the Swagger schema such that
it now properly reflects the event that is being created.

Hooray for tests that catch things like this.
........

Merged revisions 434597 from http://svn.asterisk.org/svn/asterisk/branches/13

git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@434598 65c4cc65-6c06-0410-ace0-fbb531ad65f3

4 years agodns: Fix build when TEST_FRAMEWORK is not defined.
Joshua Colp [Fri, 10 Apr 2015 13:32:24 +0000 (13:32 +0000)]
dns: Fix build when TEST_FRAMEWORK is not defined.

git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@434583 65c4cc65-6c06-0410-ace0-fbb531ad65f3

4 years agochannels/chan_iax2: Improve POKE expiration time calculation for lossy networks
Matthew Jordan [Fri, 10 Apr 2015 12:40:36 +0000 (12:40 +0000)]
channels/chan_iax2: Improve POKE expiration time calculation for lossy networks

POKE is used to check for peer availability; however, in networks with packet
loss, the current calculations may result in POKE expiration times that are too
short. This patch alters the expiration/retry time logic to take into account
the last known qualify round trip time, as opposed to always using a static
value for each peer.

Review: https://reviewboard.asterisk.org/r/4536

ASTERISK-22352 #close
Reported by: Frederic Van Espen

ASTERISK-24894 #close
Reported by: Y Ateya
patches:
  poke_noanswer_duration.diff submitted by Y Ateya (License 6693)
........

Merged revisions 434564 from http://svn.asterisk.org/svn/asterisk/branches/11
........

Merged revisions 434565 from http://svn.asterisk.org/svn/asterisk/branches/13

git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@434566 65c4cc65-6c06-0410-ace0-fbb531ad65f3

4 years agochannels/chan_iax2: Add a configuration parameter for call token expiration
Matthew Jordan [Fri, 10 Apr 2015 12:23:42 +0000 (12:23 +0000)]
channels/chan_iax2: Add a configuration parameter for call token expiration

This patch adds a new configuration parameter, 'calltokenexpiration', that
controls how long before an authentication call token is expired. The default
maintains the RFC specified 10 seconds. Setting it to a higher value may be
useful in lossy networks.

Review: https://reviewboard.asterisk.org/r/4588

ASTERISK-24939 #close
Reported by: Y Ateya
patches:
  ctoken_configuration.diff submitted by Y Ateya (License 6693)

git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@434563 65c4cc65-6c06-0410-ace0-fbb531ad65f3

4 years agores_pjsip_phoneprov_provider: Fix reference leak on unload
George Joseph [Thu, 9 Apr 2015 23:12:13 +0000 (23:12 +0000)]
res_pjsip_phoneprov_provider: Fix reference leak on unload

res_pjsip_phoneprov_provider was leaking references to phoneprov objects due to
a missing OBJ_NODATA in an ao2_callback in load_users().  Rather than adding the
OBJ_NODATA, I changed load_users to use a more straightforward ao2_iterator.
This plugged the leak but exposed an unload order issue between
res_pjsip_phoneprov_provider, res_phoneprov and res_pjsip.

res_pjsip_phoneprov_provider unloads first, then res_phoneprov, then res_pjsip.
Since res_pjsip_phoneprov_provider uses res_pjsip's sorcery instance, when it
unloads, it's objects are still in the sorcery instance.  When res_pjsip
unloads, it destroys all its objects including res_pjsip_phoneprov_provider's.
The phoneprov destructor then attempts to unregister the extension from
res_phoneprov but because res_phoneprov is already cleaned up, its users
container is gone and we get a FRACK.

Simple solution, check for the NULL users container before attempting to remove
the entry. Duh.

Ran tests/res_phoneprov/res_phoneprov_provider.  No leaks in
res_pjsip_phoneprov_provider and no FRACKs.

Reported-by: Corey Farrell
Tested-by: George Joseph
Review: https://reviewboard.asterisk.org/r/4608/
ASTERISK-24935 #close
........

Merged revisions 434545 from http://svn.asterisk.org/svn/asterisk/branches/13

git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@434547 65c4cc65-6c06-0410-ace0-fbb531ad65f3

4 years agoloader/main: Don't set ast_fully_booted until deferred reloads are processed
George Joseph [Thu, 9 Apr 2015 23:08:10 +0000 (23:08 +0000)]
loader/main: Don't set ast_fully_booted until deferred reloads are processed

Until we have a true module management facility it's sometimes necessary for one
module to force a reload on another before its own load is complete.  If
Asterisk isn't fully booted yet, these reloads are deferred.  The problem is
that asterisk reports fully booted before processing the deferred reloads which
means Asterisk really isn't quite ready when it says it is.

This patch moves the report of fully booted after the processing of the deferred
reloads is complete.

Since the pjsip stack has the most number of related modules, I ran the
channels/pjsip testsuite to make sure there aren't any issues.  All tests
passed.

Tested-by: George Joseph
Review: https://reviewboard.asterisk.org/r/4604/
........

Merged revisions 434544 from http://svn.asterisk.org/svn/asterisk/branches/13

git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@434546 65c4cc65-6c06-0410-ace0-fbb531ad65f3

4 years agores_pjsip: add CLI command to show global and system configuration
Kevin Harwell [Thu, 9 Apr 2015 22:07:50 +0000 (22:07 +0000)]
res_pjsip: add CLI command to show global and system configuration

Added a new CLI command for res_pjsip that shows both global and system
configuration settings: pjsip show settings

ASTERISK-24918 #close
Reported by: Scott Griepentrog
Review: https://reviewboard.asterisk.org/r/4597/
........

Merged revisions 434527 from http://svn.asterisk.org/svn/asterisk/branches/13

git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@434528 65c4cc65-6c06-0410-ace0-fbb531ad65f3

4 years agochan_iax2.c: Fix ref leak in iax2_request().
Richard Mudgett [Thu, 9 Apr 2015 16:09:58 +0000 (16:09 +0000)]
chan_iax2.c: Fix ref leak in iax2_request().

* Increased warning message format capability string buffer size in
iax2_request().

Review: https://reviewboard.asterisk.org/r/4601/
........

Merged revisions 434510 from http://svn.asterisk.org/svn/asterisk/branches/13

git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@434511 65c4cc65-6c06-0410-ace0-fbb531ad65f3

4 years agobridge_native_rtp.c: Defer allocation and check if it fails in native_rtp_bridge_comp...
Richard Mudgett [Thu, 9 Apr 2015 16:05:44 +0000 (16:05 +0000)]
bridge_native_rtp.c: Defer allocation and check if it fails in native_rtp_bridge_compatible().

Review: https://reviewboard.asterisk.org/r/4601/
........

Merged revisions 434508 from http://svn.asterisk.org/svn/asterisk/branches/13

git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@434509 65c4cc65-6c06-0410-ace0-fbb531ad65f3

4 years agores/res_pjsip_dlg_options: Add a module to handle in-dialog OPTIONS requests
Matthew Jordan [Thu, 9 Apr 2015 15:43:20 +0000 (15:43 +0000)]
res/res_pjsip_dlg_options: Add a module to handle in-dialog OPTIONS requests

This patch adds a new session supplement that handles in-dialog OPTIONS
requests. Said OPTIONS requests are sent a 200 OK, as an endpoint lookup
for the OPTIONS request would already have been done by the time the
session supplement receives the inbound request.

ASTERISK-24862 #close
Reported by: yaron nahum
patches:
  res_pjsip_dlg_options.c submitted by yaron nahum (License 6676)
........

Merged revisions 434506 from http://svn.asterisk.org/svn/asterisk/branches/13

git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@434507 65c4cc65-6c06-0410-ace0-fbb531ad65f3

4 years agoReduce duplication of common DNS code.
Mark Michelson [Thu, 9 Apr 2015 14:58:02 +0000 (14:58 +0000)]
Reduce duplication of common DNS code.

The NAPTR and SRV branches were worked on independently and
resulted in some code being duplicated in each. Since both
have been merged into trunk now, this patch reduces the
duplication by factoring out common code into its own
source files.

git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@434490 65c4cc65-6c06-0410-ace0-fbb531ad65f3

4 years agoclang compiler warnings: Fix autological comparisons
Matthew Jordan [Thu, 9 Apr 2015 12:57:21 +0000 (12:57 +0000)]
clang compiler warnings: Fix autological comparisons

This fixes autological comparison warnings in the following:
 * chan_skinny: letohl may return a signed or unsigned value, depending on the
   macro chosen
 * func_curl: Provide a specific cast to CURLoption to prevent mismatch
 * cel: Fix enum comparisons where the enum can never be negative
 * enum: Fix comparison of return result of dn_expand, which returns a signed
   int value
 * event: Fix enum comparisons where the enum can never be negative
 * indications: tone_data.freq1 and freq2 are unsigned, and hence can never be
   negative
 * presencestate: Use the actual enum value for INVALID state
 * security_events: Fix enum comparisons where the enum can never be negative
 * udptl: Don't bother to check if the return value from encode_length is less
   than 0, as it returns an unsigned int
 * translate: Since the parameters are unsigned int, don't bother checking
   to see if they are negative. The cast to unsigned int would already blow
   past the matrix bounds.
 * res_pjsip_exten_state: Use a temporary value to cache the return of
   ast_hint_presence_state
 * res_stasis_playback: Fix enum comparisons where the enum can never be
   negative
 * res_stasis_recording: Add an enum value for the case where the recording
   operation is in error; fix enum comparisons
 * resource_bridges: Use enum value as opposed to -1
 * resource_channels: Use enum value as opposed to -1

Review: https://reviewboard.asterisk.org/r/4533
ASTERISK-24917
Reported by: dkdegroot
patches:
  rb4533.patch submitted by dkdegroot (License 6600)
........

Merged revisions 434469 from http://svn.asterisk.org/svn/asterisk/branches/11
........

Merged revisions 434470 from http://svn.asterisk.org/svn/asterisk/branches/13

git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@434471 65c4cc65-6c06-0410-ace0-fbb531ad65f3

4 years agoapps/app_queue: Prevent possible crash when evaluating queue penalty rules
Matthew Jordan [Thu, 9 Apr 2015 02:05:26 +0000 (02:05 +0000)]
apps/app_queue: Prevent possible crash when evaluating queue penalty rules

Although it only occurred once, a crash occurred when a queue attempted to
evaluate a queue penalty rule that appeared to have already been destroyed.
In many locations in app_queue, a test is done to see if qe->pr is NULL;
however, when we dispose of a queue's penalty rules, we don't set the pointer
to NULL after free'ing it. This patch does that to prevent any dangling
pointers from lingering on the queue object.

Review: https://reviewboard.asterisk.org/r/4522

ASTERISK-23319 #close
Reported by: Vadim
patches:
  rb4552.patch submitted by Stefan Engström (License 6691)
........

Merged revisions 434448 from http://svn.asterisk.org/svn/asterisk/branches/11
........

Merged revisions 434449 from http://svn.asterisk.org/svn/asterisk/branches/13

git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@434450 65c4cc65-6c06-0410-ace0-fbb531ad65f3

4 years agores_pjsip_t38: Fix FAX failures when using PJSIP with authentication
Jonathan Rose [Wed, 8 Apr 2015 18:32:31 +0000 (18:32 +0000)]
res_pjsip_t38: Fix FAX failures when using PJSIP with authentication

Without this patch, if a PJSIP endpoint with udptl enabled and authentication
set attempted to use sendFax, the FAX session would fail during setup. This
was because the invite issued in response to being auth challenged would cause
the PJSIP channel performing the FAX to receive a second T38 framehook and
this would cause frames to be consumed in an inappropriate manner.

ASTERISK-24933 #close
Reported by: Jonathan Rose
Review: https://reviewboard.asterisk.org/r/4577/
........

Merged revisions 434425 from http://svn.asterisk.org/svn/asterisk/branches/13

git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@434431 65c4cc65-6c06-0410-ace0-fbb531ad65f3

4 years agoBridging: Eliminate the unnecessary make channel compatible with bridge operation.
Richard Mudgett [Wed, 8 Apr 2015 18:20:26 +0000 (18:20 +0000)]
Bridging: Eliminate the unnecessary make channel compatible with bridge operation.

When a channel enters the bridging system it is first made compatible with
the bridge and then the bridge technology makes the channel compatible
with the technology.  For all but the DAHDI native and softmix bridge
technologies the make channel compatible with the bridge step is an
effective noop because the other technologies allow all audio formats.
For the DAHDI native bridge technology it doesn't matter because it is not
an initial bridge technology and chan_dahdi allows only one native format
per channel.  For the softmix bridge technology, it is a noop at best and
harmful at worst because the wrong translation path could be setup if the
channel's native formats allow more than one audio format.

This is an intermediate patch for a series of patches aimed at improving
translation path choices.

* Removed code dealing with the unnecessary step of making the channel
compatible with the bridge.

ASTERISK-24841
Reported by: Matt Jordan

Review: https://reviewboard.asterisk.org/r/4600/
........

Merged revisions 434424 from http://svn.asterisk.org/svn/asterisk/branches/13

git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@434430 65c4cc65-6c06-0410-ace0-fbb531ad65f3

4 years agoSecurity/tcptls: MitM Attack potential from certificate with NULL byte in CN.
Jonathan Rose [Wed, 8 Apr 2015 16:49:18 +0000 (16:49 +0000)]
Security/tcptls: MitM Attack potential from certificate with NULL byte in CN.

When registering to a SIP server with TLS, Asterisk will accept CA signed
certificates with a common name that was signed for a domain other than the
one requested if it contains a null character in the common name portion of
the cert. This patch fixes that by checking that the common name length
matches the the length of the content we actually read from the common name
segment. Some certificate authorities automatically sign CA requests when
the requesting CN isn't already taken, so an attacker could potentially
register a CN with something like www.google.com\x00www.secretlyevil.net
and have their certificate signed and Asterisk would accept that certificate
as though it had been for www.google.com - this is a security fix and is
noted in AST-2015-003.

ASTERISK-24847 #close
Reported by: Maciej Szmigiero
Patches:
 asterisk-null-in-cn.patch submitted by mhej (license 6085)
........

Merged revisions 434337 from http://svn.asterisk.org/svn/asterisk/branches/1.8
........

Merged revisions 434338 from http://svn.asterisk.org/svn/asterisk/branches/11
........

Merged revisions 434384 from http://svn.asterisk.org/svn/asterisk/branches/13

git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@434385 65c4cc65-6c06-0410-ace0-fbb531ad65f3

4 years agoformat_cache.c: Add missing slin12 format to ast_format_cache_is_slinear().
Richard Mudgett [Wed, 8 Apr 2015 16:31:45 +0000 (16:31 +0000)]
format_cache.c: Add missing slin12 format to ast_format_cache_is_slinear().
........

Merged revisions 434357 from http://svn.asterisk.org/svn/asterisk/branches/13

git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@434383 65c4cc65-6c06-0410-ace0-fbb531ad65f3

4 years agochan_iax2: Fix compilation issue due to funky merge
Matthew Jordan [Wed, 8 Apr 2015 12:02:37 +0000 (12:02 +0000)]
chan_iax2: Fix compilation issue due to funky merge

Don't mix declarations and code!

git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@434294 65c4cc65-6c06-0410-ace0-fbb531ad65f3

4 years agochan_iax2: Fix crash caused by unprotected access to iaxs[peer->callno]
Matthew Jordan [Wed, 8 Apr 2015 12:00:35 +0000 (12:00 +0000)]
chan_iax2: Fix crash caused by unprotected access to iaxs[peer->callno]

This patch fixes an access to the peer callnumber that is unprotected by a
corresponding mutex. The peer->callno value can be changed by multiple threads,
and all data inside the iaxs array must be procted by a corresponding lock
of iaxsl.

The patch moves the unprotected access to a location where the mutex is
safely obtained.

Review: https://reviewboard.asterisk.org/r/4599/

ASTERISK-21211 #close
Reported by: Jaco Kroon
patches:
  asterisk-11.2.1-iax2_poke-segfault.diff submitted by Jaco Kroon (License 5671)
........

Merged revisions 434291 from http://svn.asterisk.org/svn/asterisk/branches/11
........

Merged revisions 434292 from http://svn.asterisk.org/svn/asterisk/branches/13

git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@434293 65c4cc65-6c06-0410-ace0-fbb531ad65f3

4 years agochan_sip: Handle IPv4 mapped IPv6 clients when NAT is enabled
Matthew Jordan [Wed, 8 Apr 2015 11:54:09 +0000 (11:54 +0000)]
chan_sip: Handle IPv4 mapped IPv6 clients when NAT is enabled

When udpbindaddr is set to the IPv6 bind all address of '::', Asterisk will
attempt to handle both IPv4 and IPv6 addresses, although the information will
be stored in a struct with an AF_INET6 address type. However, the current
NAT handling code won't handle the IPv4 mapped IPv6 addresses correctly.
This patch adds an additional check for the mapped address case, allowing
the NAT code to handle clients even when the address is IPv6.

Review: https://reviewboard.asterisk.org/r/4563/

ASTERISK-18032 #close
Reported by: Christoph Timm
patches:
  nat_with_ipv6.diff submitted by Valentin Vidić (License 6697)
........

Merged revisions 434288 from http://svn.asterisk.org/svn/asterisk/branches/11
........

Merged revisions 434289 from http://svn.asterisk.org/svn/asterisk/branches/13

git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@434290 65c4cc65-6c06-0410-ace0-fbb531ad65f3

4 years agoclang compiler warnings: Fix pointer-bool-converesion warnings
Matthew Jordan [Wed, 8 Apr 2015 11:45:05 +0000 (11:45 +0000)]
clang compiler warnings: Fix pointer-bool-converesion warnings

This patch fixes several warnings pointed out by the clang compiler.
* chan_pjsip: Removed check for data->text, as it will always be non-NULL.
* app_minivm: Fixed evaluation of etemplate->locale, which will always
  evaluate to 'true'. This patch changes the evaluation to use
  ast_strlen_zero.
* app_queue:
  - Fixed evaluation of qe->parent->monfmt, which always evaluates to
    true. Instead, we just check to see if the dereferenced pointer
    evaluates to true.
  - Fixed evaluation of mem->state_interface, wrapping it with a call to
    ast_strlen_zero.
* res_smdi: Wrapped search_msg->mesg_desk_term with calls to ast_strlen_zero.

Review: https://reviewboard.asterisk.org/r/4541

ASTERISK-24917
Reported by: dkdegroot
patches:
  rb4541.patch submitted by dkdegroot (License 6600)
........

Merged revisions 434285 from http://svn.asterisk.org/svn/asterisk/branches/11
........

Merged revisions 434286 from http://svn.asterisk.org/svn/asterisk/branches/13

git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@434287 65c4cc65-6c06-0410-ace0-fbb531ad65f3

4 years agocel_pgsl: Add support for GMT timestamps
Matthew Jordan [Wed, 8 Apr 2015 11:35:53 +0000 (11:35 +0000)]
cel_pgsl: Add support for GMT timestamps

This patch adds a new option to cel_pgsl, "usegmtime", which causes timestamps
to be logged in GMT.

Review: https://reviewboard.asterisk.org/r/4571/

ASTERISK-23186 #close
Reported by: Rodrigo Ramirez Norambuena
patches:
  cel_pgsql.c_add_usegmtime2.patch submitted by Rodrigo Ramirez Norambuena (License 6577)

git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@434284 65c4cc65-6c06-0410-ace0-fbb531ad65f3

4 years agopjsip: resolve compatibility problem with ast_sip_session
Scott Griepentrog [Tue, 7 Apr 2015 19:40:22 +0000 (19:40 +0000)]
pjsip: resolve compatibility problem with ast_sip_session

A change in r430179 inserted a variable near the top of a
structure caused a problem when running DPMA in a version
of Asterisk compiled across the change.  This patch moves
the new variable to the end of the structure, eliminating
the problem.

Review: https://reviewboard.asterisk.org/r/4574/
........

Merged revisions 433944 from http://svn.asterisk.org/svn/asterisk/branches/13
........

Merged revisions 434261 from http://svn.asterisk.org/svn/asterisk/branches/13

git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@434263 65c4cc65-6c06-0410-ace0-fbb531ad65f3

4 years agobridge.c: Hangup attended transfer target after it has been swapped out
Kevin Harwell [Tue, 7 Apr 2015 16:42:18 +0000 (16:42 +0000)]
bridge.c: Hangup attended transfer target after it has been swapped out

After completing an attended transfer the transfer target channel (the one that
gets swapped out) was not being hung up after leaving the bridge. This resulted
in a channel possibly being left around. Added an explicit softhangup for the
channel in question after the transfer is successfully completed in order to
make sure the channel is hung up.

ASTERISK-24782 #close
Reported by: John Bigelow
Review: https://reviewboard.asterisk.org/r/4575/
........

Merged revisions 434240 from http://svn.asterisk.org/svn/asterisk/branches/13

git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@434241 65c4cc65-6c06-0410-ace0-fbb531ad65f3

4 years agoDo not queue message requests that we do not respond to.
Mark Michelson [Tue, 7 Apr 2015 15:34:53 +0000 (15:34 +0000)]
Do not queue message requests that we do not respond to.

If we receive a MESSAGE request that we cannot send a response
to, we should not send the incoming MESSAGE to the dialplan.

This commit should help the bouncing message_retrans test to
pass consistently.
........

Merged revisions 434218 from http://svn.asterisk.org/svn/asterisk/branches/13

git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@434219 65c4cc65-6c06-0410-ace0-fbb531ad65f3

4 years agoARI: Add the ability to intercept hold and raise an event
Matthew Jordan [Tue, 7 Apr 2015 15:22:42 +0000 (15:22 +0000)]
ARI: Add the ability to intercept hold and raise an event

For some applications - such as SLA - a phone pressing hold should not behave
in the fashion that the Asterisk core would like it to. Instead, the hold
action has some application specific behaviour associated with it - such as
disconnecting the channel that initiated the hold; only playing MoH to channels
in the bridge if the channels are of a particular type, etc.

One way of accomplishing this is to use a framehook to intercept the
hold/unhold frames, raise an event, and eat the frame. Tasty. This patch
accomplishes that using a new dialplan function, HOLD_INTERCEPT.

In addition, some general cleanup of raising hold/unhold Stasis messages was
done, including removing some RAII_VAR usage.

Review: https://reviewboard.asterisk.org/r/4549/

ASTERISK-24922 #close
........

Merged revisions 434216 from http://svn.asterisk.org/svn/asterisk/branches/13

git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@434217 65c4cc65-6c06-0410-ace0-fbb531ad65f3

4 years agoclang compiler warnings: Fix sometimes-initialized warning in func_math
Matthew Jordan [Tue, 7 Apr 2015 02:10:31 +0000 (02:10 +0000)]
clang compiler warnings: Fix sometimes-initialized warning in func_math

This patch fixes a bug in a unit test in func_math where a variable could be
passed to ast_free that wasn't allocated. This patch corrects the issue and
ensures that we only attempt to free a variable if we previously allocated
it.

Review: https://reviewboard.asterisk.org/r/4552

ASTERISK-24917
Reported by: dkdegroot
patches:
  rb4552.patch submitted by dkdegroot (License 6600)
........

Merged revisions 434190 from http://svn.asterisk.org/svn/asterisk/branches/11
........

Merged revisions 434191 from http://svn.asterisk.org/svn/asterisk/branches/13

git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@434192 65c4cc65-6c06-0410-ace0-fbb531ad65f3

4 years agoclang compiler warnings: Fix non-literal-null-conversion warnings
Matthew Jordan [Tue, 7 Apr 2015 02:03:48 +0000 (02:03 +0000)]
clang compiler warnings: Fix non-literal-null-conversion warnings

Clang will flag errors when a char pointer is set to '\0', as opposed to a
value that the char pointer points to. This patch fixes this warning
in a variety of locations.

Review: https://reviewboard.asterisk.org/r/4551

ASTERISK-24917
Reported by: dkdegroot
patches:
  rb4551.patch submitted by dkdegroot (License 6600)
........

Merged revisions 434187 from http://svn.asterisk.org/svn/asterisk/branches/11
........

Merged revisions 434188 from http://svn.asterisk.org/svn/asterisk/branches/13

git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@434189 65c4cc65-6c06-0410-ace0-fbb531ad65f3

4 years agoUncomment test case.
Mark Michelson [Mon, 6 Apr 2015 21:54:53 +0000 (21:54 +0000)]
Uncomment test case.

git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@434170 65c4cc65-6c06-0410-ace0-fbb531ad65f3

4 years agoAdd missing DNS NAPTR test file.
Mark Michelson [Mon, 6 Apr 2015 21:13:36 +0000 (21:13 +0000)]
Add missing DNS NAPTR test file.

git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@434154 65c4cc65-6c06-0410-ace0-fbb531ad65f3

4 years agores_pjsip: config option 'timers' can't be set to 'no'
Kevin Harwell [Mon, 6 Apr 2015 19:23:57 +0000 (19:23 +0000)]
res_pjsip: config option 'timers' can't be set to 'no'

When setting the configuration option 'timers' equal to 'no' the bit flag was
not properly negated. This patch clears all associated flags and only sets the
specified one. pjsip will handle any necessary flag combinations. Also went
ahead and did similar for the '100rel' option.

ASTERISK-24910 #close
Reported by: Ray Crumrine
Review: https://reviewboard.asterisk.org/r/4582/
........

Merged revisions 434131 from http://svn.asterisk.org/svn/asterisk/branches/13

git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@434132 65c4cc65-6c06-0410-ace0-fbb531ad65f3

4 years agobuild: Fixes for gcc 5 compilation
George Joseph [Mon, 6 Apr 2015 19:04:32 +0000 (19:04 +0000)]
build: Fixes for gcc 5 compilation

These are fixes for compilation under gcc 5.0...

chan_sip.c:    In parse_request needed to make 'lim' unsigned.
inline_api.h:  Needed to add a check for '__GNUC_STDC_INLINE__' to detect C99
               inline semantics (same as clang).
ccss.c:        In ast_cc_set_parm, needed to fix weird comparison.
dsp.c:         Needed to work around a possible compiler bug.  It was throwing
               an array-bounds error but neither
               sgriepentrog, rmudgett nor I could figure out why.
manager.c:     In action_atxfer, needed to correct an array allocation.

This patch will go to 11, 13, trunk.

Review: https://reviewboard.asterisk.org/r/4581/
Reported-by: Jeffrey Ollie
Tested-by: George Joseph
ASTERISK-24932 #close
........

Merged revisions 434113 from http://svn.asterisk.org/svn/asterisk/branches/11
........

Merged revisions 434114 from http://svn.asterisk.org/svn/asterisk/branches/13

git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@434115 65c4cc65-6c06-0410-ace0-fbb531ad65f3

4 years agoclang compiler warnings: Remove large chunks of unused code from extconf
Matthew Jordan [Mon, 6 Apr 2015 18:18:32 +0000 (18:18 +0000)]
clang compiler warnings: Remove large chunks of unused code from extconf

This patch fixes a warning caught by clang, in which it detected that large
chunks of extconf were unused. Frankly, I wish we could pretend that all of
extconf was unused, but alas, that is not yet the case.

A few extraneous functions in the parking tests were removed as well, for
the same reason.

Review: https://reviewboard.asterisk.org/r/4553

ASTERISK-24917
Reported by: dkdegroot
patches:
  rb4553.patch submitted by dkdegroot (License 6600)
........

Merged revisions 434093 from http://svn.asterisk.org/svn/asterisk/branches/11
........

Merged revisions 434097 from http://svn.asterisk.org/svn/asterisk/branches/13

git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@434099 65c4cc65-6c06-0410-ace0-fbb531ad65f3

4 years agoclang compiler warnings: Fix sometimes-uninitialized warning in pbx_config
Matthew Jordan [Mon, 6 Apr 2015 18:03:54 +0000 (18:03 +0000)]
clang compiler warnings: Fix sometimes-uninitialized warning in pbx_config

This patch fixes a warning caught by clang, in which a char pointer could be
assigned to before it was initialized. The patch re-organizes the code to
ensure that the pointer is always initialized, even on off nominal paths.

Review: https://reviewboard.asterisk.org/r/4529

ASTERISK-24917
Reported by: dkdegroot
patches:
  rb4529.patch submitted by dkdegroot (License 6600)
........

Merged revisions 434090 from http://svn.asterisk.org/svn/asterisk/branches/11
........

Merged revisions 434091 from http://svn.asterisk.org/svn/asterisk/branches/13

git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@434092 65c4cc65-6c06-0410-ace0-fbb531ad65f3

4 years agoclang compiler warnings: Fix format specified in framehook
Matthew Jordan [Mon, 6 Apr 2015 17:52:31 +0000 (17:52 +0000)]
clang compiler warnings: Fix format specified in framehook

This patch fixes an invalid format specifier used in the formatting of an
ERROR message in the framehook code. The format specifier specifies a
type of 'unsigned short', but the argument passed to it is of type 'int'.
The patch changes the format specifier to 'i'.

Review: https://reviewboard.asterisk.org/r/4540

ASTERISK-24917
Reported by: dkdegroot
patches:
  rb4535.patch submitted by dkdegroot (License 6600)
........

Merged revisions 434087 from http://svn.asterisk.org/svn/asterisk/branches/11
........

Merged revisions 434088 from http://svn.asterisk.org/svn/asterisk/branches/13

git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@434089 65c4cc65-6c06-0410-ace0-fbb531ad65f3

4 years agoMerge NAPTR support into trunk.
Mark Michelson [Mon, 6 Apr 2015 17:05:47 +0000 (17:05 +0000)]
Merge NAPTR support into trunk.

This adds NAPTR record allocation and sorting, as well as
unit tests that verify that NAPTR records are parsed and
sorted correctly.

Review: https://reviewboard.asterisk.org/r/4542

git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@434068 65c4cc65-6c06-0410-ace0-fbb531ad65f3

4 years agoEnsure that a non-zero sample rate is returned for all formats.
Mark Michelson [Mon, 6 Apr 2015 16:02:28 +0000 (16:02 +0000)]
Ensure that a non-zero sample rate is returned for all formats.

Versions of Asterisk prior to 12 defaulted to 8000 as a sample rate
if one was not provided by a format. In Asterisk 13, this was removed.
The result was that some calculations which involve dividing by the
sample rate resulted in dividing by 0. The fix being put in place
here is to have the same default fallback that was present in previous
versions of Asterisk.

Asterisk-24914 #close
Reported by Marcello Ceschia
........

Merged revisions 434046 from http://svn.asterisk.org/svn/asterisk/branches/13

git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@434047 65c4cc65-6c06-0410-ace0-fbb531ad65f3

4 years agores_pjsip_phoneprov_provider: Revert 433996 / 433997.
Corey Farrell [Mon, 6 Apr 2015 15:17:18 +0000 (15:17 +0000)]
res_pjsip_phoneprov_provider: Revert 433996 / 433997.

res_pjsip_phoneprov_provider is using ao2_callback with OBJ_MULTIPLE, then
ignoring the return.  OBJ_NODATA flag was to prevent a reference leak, but
this caused the module to FRACK on unload.  Revert change until this can
be investigated further.

ASTERISK-24935
Reported by: Corey Farrell
Review: https://reviewboard.asterisk.org/r/4578/
........

Merged revisions 434025 from http://svn.asterisk.org/svn/asterisk/branches/13

git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@434026 65c4cc65-6c06-0410-ace0-fbb531ad65f3

4 years agoParkedCall: Don't allow dialplan fallthrough after retrieving parked call.
Mark Michelson [Mon, 6 Apr 2015 14:51:21 +0000 (14:51 +0000)]
ParkedCall: Don't allow dialplan fallthrough after retrieving parked call.

This is a change to align behavior with that of Asterisk 11 and previous versions.
In those versions, if a parked call were retrieved, and the call ended, the parked
call retriever would be hung up after the ParkedCall application ran. Prior to this
patch, in Asterisk 13, the same situation would result in the parked call retriever
falling through to additional priorities in the extension where the ParkedCall
application was called. With this patch, the behavior between Asterisk 11 and 13
aligns.

ASTERISK-24899 #close
Reported by Malcolm Davenport
Patches:
ASTERISK-24899.patch uploaded by Mark Michelson(license #5049)
........

Merged revisions 434022 from http://svn.asterisk.org/svn/asterisk/branches/13

git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@434023 65c4cc65-6c06-0410-ace0-fbb531ad65f3

4 years agores_pjsip_phoneprov_provider: Fix leaked OBJ_MULTIPLE iterator.
Corey Farrell [Sun, 5 Apr 2015 12:55:21 +0000 (12:55 +0000)]
res_pjsip_phoneprov_provider: Fix leaked OBJ_MULTIPLE iterator.

res_pjsip_phoneprov_provider was using ao2_callback with OBJ_MULTIPLE, then
ignoring the return.  Added OBJ_NODATA flag to prevent a reference leak.

ASTERISK-24935 #close
Reported by: Corey Farrell
Review: https://reviewboard.asterisk.org/r/4578/
........

Merged revisions 433996 from http://svn.asterisk.org/svn/asterisk/branches/13

git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@433997 65c4cc65-6c06-0410-ace0-fbb531ad65f3

4 years agores_pjsip_messaging: Serialize outbound SIP MESSAGEs
Mark Michelson [Fri, 3 Apr 2015 21:54:36 +0000 (21:54 +0000)]
res_pjsip_messaging: Serialize outbound SIP MESSAGEs

Outbound SIP MESSAGEs had the potential to be sent out
of order from how they were specified in a set of
dialplan steps.

This change creates a serializer for sending outbound
MESSAGE requests on. This ensures that the MESSAGEs are
sent by Asterisk in the same order that they were sent
from the dialplan.

ASTERISK-24937 #close
Reported by Mark Michelson

Review: https://reviewboard.asterisk.org/r/4579
........

Merged revisions 433968 from http://svn.asterisk.org/svn/asterisk/branches/13

git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@433969 65c4cc65-6c06-0410-ace0-fbb531ad65f3

4 years agopjsip: resolve compatibility problem with ast_sip_session
Scott Griepentrog [Thu, 2 Apr 2015 14:56:39 +0000 (14:56 +0000)]
pjsip: resolve compatibility problem with ast_sip_session

A change in r430179 inserted a variable near the top of a
structure caused a problem when running DPMA in a version
of Asterisk compiled across the change.  This patch moves
the new variable to the end of the structure, eliminating
the problem.

Review: https://reviewboard.asterisk.org/r/4574/
........

Merged revisions 433944 from http://svn.asterisk.org/svn/asterisk/branches/13

git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@433945 65c4cc65-6c06-0410-ace0-fbb531ad65f3

4 years agoTell menuselect that MALLOC_DEBUG conflicts with DEBUG_CHAOS.
Corey Farrell [Thu, 2 Apr 2015 10:38:03 +0000 (10:38 +0000)]
Tell menuselect that MALLOC_DEBUG conflicts with DEBUG_CHAOS.

DEBUG_CHAOS was marked as conflicting with MALLOC_DEBUG, but
for this to work correctly MALLOC_DEBUG must also be marked
as conflicting with DEBUG_CHAOS.

Review: https://reviewboard.asterisk.org/r/4557/
........

Merged revisions 433923 from http://svn.asterisk.org/svn/asterisk/branches/13

git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@433924 65c4cc65-6c06-0410-ace0-fbb531ad65f3

4 years agostasis: set a channel variable on websocket disconnect error
Ashley Sanders [Wed, 1 Apr 2015 16:30:25 +0000 (16:30 +0000)]
stasis: set a channel variable on websocket disconnect error

Resolve compile errors caused by r433863 by fixing the
documentation xml to comply with the schema.
........

Merged revisions 433888 from http://svn.asterisk.org/svn/asterisk/branches/13

git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@433891 65c4cc65-6c06-0410-ace0-fbb531ad65f3

4 years agodns: Add support for SRV record parsing and sorting.
Joshua Colp [Wed, 1 Apr 2015 16:27:48 +0000 (16:27 +0000)]
dns: Add support for SRV record parsing and sorting.

This change adds support for parsing SRV records and consuming their values
in an easy fashion. It also adds automatic sorting of SRV records according
to RFC 2782.

Tests have also been included which cover parsing, sorting, and off-nominal
cases where the record is corrupted.

ASTERISK-24931 #close
Reported by: Joshua Colp

Review: https://reviewboard.asterisk.org/r/4528/

git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@433889 65c4cc65-6c06-0410-ace0-fbb531ad65f3

4 years agostasis: set a channel variable on websocket disconnect error
Mark Michelson [Wed, 1 Apr 2015 13:35:10 +0000 (13:35 +0000)]
stasis: set a channel variable on websocket disconnect error

Resolve compile errors caused by r433839 by included the missing
header file, pbx.h.
........

Merged revisions 433863 from http://svn.asterisk.org/svn/asterisk/branches/13

git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@433868 65c4cc65-6c06-0410-ace0-fbb531ad65f3

4 years agostasis: set a channel variable on websocket disconnect error
Ashley Sanders [Tue, 31 Mar 2015 22:49:45 +0000 (22:49 +0000)]
stasis: set a channel variable on websocket disconnect error

When an error occurs while writing to a web socket, the web socket is
disconnected and the event is logged. A side-effect of this, however, is that
any application on the other side waiting for a response from Stasis is left
hanging indefinitely (as there is no mechanism presently available for
notifying interested parties about web socket error states in Stasis).

To remedy this scenario, this patch introduces a new channel variable:
STASISSTATUS.

The possible values for STASISSTATUS are:
SUCCESS         - The channel has exited Stasis without any failures
FAILED          - Something caused Stasis to croak. Some (not all) possible
                  reasons for this:
                    - The app registry is not instantiated;
                    - The app requested is not registered;
                    - The app requested is not active;
                    - Stasis couldn't send a start message

ASTERISK-24802
Reported By: Kevin Harwell
Review: https://reviewboard.asterisk.org/r/4519/
........

Merged revisions 433839 from http://svn.asterisk.org/svn/asterisk/branches/13

git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@433845 65c4cc65-6c06-0410-ace0-fbb531ad65f3

4 years agochan_sip: Fix expression in unit test /channels/chan_sip/test_sip_rtpqos.
Richard Mudgett [Tue, 31 Mar 2015 17:04:18 +0000 (17:04 +0000)]
chan_sip: Fix expression in unit test /channels/chan_sip/test_sip_rtpqos.

Fix misplaced parentheses in original fabs() expression.
........

Merged revisions 433816 from http://svn.asterisk.org/svn/asterisk/branches/11
........

Merged revisions 433817 from http://svn.asterisk.org/svn/asterisk/branches/13

git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@433818 65c4cc65-6c06-0410-ace0-fbb531ad65f3

4 years agoBlocked revisions 433795
Corey Farrell [Tue, 31 Mar 2015 11:55:29 +0000 (11:55 +0000)]
Blocked revisions 433795

........
Re-add _ast_mem_backtrace_buffer variable for ABI compatibility.

Modules built prior to commit of r4502 expect to link at runtime
to the variable _ast_mem_backtrace_buffer.  This change re-adds
the variable to the C file only.

Review: https://reviewboard.asterisk.org/r/4558/

git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@433796 65c4cc65-6c06-0410-ace0-fbb531ad65f3

4 years agoFix an ABI compatibility issue with ast_log_safe for modules.
Corey Farrell [Mon, 30 Mar 2015 11:43:19 +0000 (11:43 +0000)]
Fix an ABI compatibility issue with ast_log_safe for modules.

Binary modules are sometimes built against the latest release of
Asterisk in each branch, and need to be compatible with all
releases of that branch.  This change ensures that utils.h only
uses ast_log_safe from the core.  For modules and utilities ast_log
is used instead.

Review: https://reviewboard.asterisk.org/r/4548/
........

Merged revisions 433772 from http://svn.asterisk.org/svn/asterisk/branches/11
........

Merged revisions 433773 from http://svn.asterisk.org/svn/asterisk/branches/13

git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@433774 65c4cc65-6c06-0410-ace0-fbb531ad65f3

4 years agoclang compiler warnings: Fix -Wabsolute-value warnings
Matthew Jordan [Mon, 30 Mar 2015 02:45:29 +0000 (02:45 +0000)]
clang compiler warnings: Fix -Wabsolute-value warnings

This patch fixes several warnings caught by clang - in this case, usage of the
abs function on non-integer values. This patch uses labs and fabs, as
appropriate, in the various affected files.

Review: https://reviewboard.asterisk.org/r/4525

ASTERISK-24917
Reported by: dkdegroot
patches:
  rb4525.patch submitted by dkdegroot (License 6600)
........

Merged revisions 433749 from http://svn.asterisk.org/svn/asterisk/branches/11
........

Merged revisions 433750 from http://svn.asterisk.org/svn/asterisk/branches/13

git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@433751 65c4cc65-6c06-0410-ace0-fbb531ad65f3

4 years agoclang compiler warnings: Fix invalid enum conversion
Matthew Jordan [Mon, 30 Mar 2015 02:39:57 +0000 (02:39 +0000)]
clang compiler warnings: Fix invalid enum conversion

This patch fixes some invalid enum conversion warnings caught by clang. In
particular:
* chan_sip: Several functions mixed usage of the st_refresher_param
  enum and st_refresher enum. This patch corrects the functions to use the
  right enum.
* chan_pjsip: Fixed mixed usage of ast_sip_session_t38state and ast_t38_state.
* strings: Fixed incorrect usage of AO2 flags with strings container.
* res_stasis: Change a return enumeration to stasis_app_user_event_res.

Review: https://reviewboard.asterisk.org/r/4535

ASTERISK-24917
Reported by: dkdegroot
patches:
  rb4535.patch submitted by dkdegroot (License 6600)
........

Merged revisions 433746 from http://svn.asterisk.org/svn/asterisk/branches/11
........

Merged revisions 433747 from http://svn.asterisk.org/svn/asterisk/branches/13

git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@433748 65c4cc65-6c06-0410-ace0-fbb531ad65f3

4 years agomain/stdtime/localtime: Fix warning introduced in r433720
Matthew Jordan [Mon, 30 Mar 2015 02:29:39 +0000 (02:29 +0000)]
main/stdtime/localtime: Fix warning introduced in r433720

The patch in r433720 caused a warning to be kicked back by gcc. It occurred
due to this check in unistd.h:

    if (__nbytes > __bos0 (__buf))
        return __read_chk_warn (__fd, __buf, __nbytes, __bos0 (__buf));

That is, if __nbytes is greater than the result of GCC's built-in object size
for the struct, we'll kick back a warning.

As it turns out, this is because there is an error in the code in the patch.
We are passing the address of the pointer to the struct, not iev, which is a
pointer to the struct. Hence, the number of bytes is probably going to be lot
larger than the number of bytes that make up a pointer! This patch changes
the code just read from the pointer to the struct - which fixes the warning.

ASTERISK-24917
........

Merged revisions 433743 from http://svn.asterisk.org/svn/asterisk/branches/11
........

Merged revisions 433744 from http://svn.asterisk.org/svn/asterisk/branches/13

git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@433745 65c4cc65-6c06-0410-ace0-fbb531ad65f3

4 years agoclang compiler warnings: Ignore -Wunused-command-line-argument
Matthew Jordan [Mon, 30 Mar 2015 01:57:07 +0000 (01:57 +0000)]
clang compiler warnings: Ignore -Wunused-command-line-argument

Asterisk's build system has a tendency to pass include directives for libraries
to everything compiled within a particular group of source files. This means
we pass the header for libxml2 to things that don't necessarily need it. As a
result, we ignore this particular warning.

Review: https://reviewboard.asterisk.org/r/4545/

ASTERISK-24917
Reported by: dkdegroot
patches:
  rb4545.patch submitted by dkdegroot (License 6600)
........

Merged revisions 433720 from http://svn.asterisk.org/svn/asterisk/branches/11
........

Merged revisions 433721 from http://svn.asterisk.org/svn/asterisk/branches/13

git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@433722 65c4cc65-6c06-0410-ace0-fbb531ad65f3

4 years agoclang compiler warnings: Fix warning for -Wgnu-variable-sized-type-not-at-end
Matthew Jordan [Mon, 30 Mar 2015 01:53:15 +0000 (01:53 +0000)]
clang compiler warnings: Fix warning for -Wgnu-variable-sized-type-not-at-end

This patch fixes a warning caught by clang, wherein a variable sized struct is
not located at the end of a struct. While the code in question actually
expected this, this is a good warning to watch for. Hence, this patch refactors
the code in question to not have two variable length elements in the same
struct.

Review: https://reviewboard.asterisk.org/r/4530/

ASTERISK-24917
Reported by: dkdegroot
patches:
  rb4530.patch submitted by dkdegroot (License 6600)
........

Merged revisions 433717 from http://svn.asterisk.org/svn/asterisk/branches/11
........

Merged revisions 433718 from http://svn.asterisk.org/svn/asterisk/branches/13

git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@433719 65c4cc65-6c06-0410-ace0-fbb531ad65f3

4 years agoclang compiler warnings: Fix a variety of "unused" warnings
Matthew Jordan [Sat, 28 Mar 2015 12:56:43 +0000 (12:56 +0000)]
clang compiler warnings: Fix a variety of "unused" warnings

This patch fixes the -Wunused-value -Wunused-variable -Wunused-const-variable
errors caught by clang. Specifically:

* apps/app_queue.c: removed unused qpm_cmd_usage[], qum_cmd_usage[],
                    qsmp_cmd_usage[]
* cel/cel_sqlite3_custom.c: removed unused name[] = "cel_sqlite3_custom"
* channels/chan_pjsip.c: removed unused desc[] = "PJSIP Channel"
* codecs/gsm/src/gsm_create.c: removed unused ident[] = "$Header$"
* funcs/func_env.c:729: Fixed ast_str_append_substr.
* main/editline/np/strlcat.c: removed unused rcsid variable
* main/editline/np/strlcpy.c: removed unused rcsid variable
* main/security_events.c: removed unused TIMESTAMP_STR_LEN
* utils/conf2ael.c: removed unused cfextension_states
* utils/extconf.c: removed unused cfextension_states

Review: https://reviewboard.asterisk.org/r/4526

ASTERISK-24917
Reported by: dkdegroot
patches:
  rb4526.patch submitted by dkdegroot (License 6600)
........

Merged revisions 433693 from http://svn.asterisk.org/svn/asterisk/branches/11
........

Merged revisions 433694 from http://svn.asterisk.org/svn/asterisk/branches/13

git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@433695 65c4cc65-6c06-0410-ace0-fbb531ad65f3

4 years agoclang compiler warnings: Fix -Wself-assign
Matthew Jordan [Sat, 28 Mar 2015 12:48:45 +0000 (12:48 +0000)]
clang compiler warnings: Fix -Wself-assign

Assigning a variable to itself isn't super useful. However, the WAV format
modules make use of this in order to perform byte endian checks. This patch
works around the warning by only performing the self assignment if we are
going to do more than just assign it to ourselves. Which is odd, but true.

Review: https://reviewboard.asterisk.org/r/4544/

ASTERISK-24917
Reported by: dkdegroot
patches:
  rb4544.patch submitted by dkdegroot (License 6600)
........

Merged revisions 433690 from http://svn.asterisk.org/svn/asterisk/branches/11
........

Merged revisions 433691 from http://svn.asterisk.org/svn/asterisk/branches/13

git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@433692 65c4cc65-6c06-0410-ace0-fbb531ad65f3

4 years agoclang compiler warnings: Fix -Wparantheses-equality warnings
Matthew Jordan [Sat, 28 Mar 2015 12:41:24 +0000 (12:41 +0000)]
clang compiler warnings: Fix -Wparantheses-equality warnings

Clang will treat ((a == b)) as a warning, as it reasonably expects that the
developer may have intended to write (a == b) or ((a = b)). This patch cleans
up all instances where equality, not assignment, was intended between two
parantheses.

Review: https://reviewboard.asterisk.org/r/4531/

ASTERISK-24917
Repoted by: dkdegroot
patches:
  rb4531.patch submitted by dkdegroot (License 6600)
........

Merged revisions 433687 from http://svn.asterisk.org/svn/asterisk/branches/11
........

Merged revisions 433688 from http://svn.asterisk.org/svn/asterisk/branches/13

git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@433689 65c4cc65-6c06-0410-ace0-fbb531ad65f3

4 years agoclang compiler warnings: Fix -Wbitfield-constant-conversion warning
Matthew Jordan [Sat, 28 Mar 2015 12:33:21 +0000 (12:33 +0000)]
clang compiler warnings: Fix -Wbitfield-constant-conversion warning

In chan_iax2, we attempt to assign a -1 to a bitfield. This gets caught by
clang, as it will truncate the -1 to a 1 implicitly.

Instead, we just assign the value a '1'.

Review: https://reviewboard.asterisk.org/r/4537/

ASTERISK-24917
Reported by: dkdegroot
patches:
  rb4537.patch submitted by dkdegroot (License 6600)
........

Merged revisions 433683 from http://svn.asterisk.org/svn/asterisk/branches/11
........

Merged revisions 433684 from http://svn.asterisk.org/svn/asterisk/branches/13

git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@433686 65c4cc65-6c06-0410-ace0-fbb531ad65f3

4 years agoclang compiler warnings: Fix -Winitializer-overrides
Matthew Jordan [Sat, 28 Mar 2015 12:32:42 +0000 (12:32 +0000)]
clang compiler warnings: Fix -Winitializer-overrides

This patch fixes clange compiler warnings for initializer overrides.
Specifically:

res_pjsip/config_transport maps PJSIP_TLSV1_METHOD to the same enumeration
value as PJSIP_SSL_DEFAULT_METHOD. When initializing an array containing
those enum values, we therefore initialize the value twice to two different
values, "tlsv1" and "default". This patch changes it to just initialize
the index in the array to "tlsv1".

Review: https://reviewboard.asterisk.org/r/4539/

ASTERISK-24917
Reported by: dkdegroot
patches:
  rb4539.patch submitted by dkdegroot (License 6600)
........

Merged revisions 433682 from http://svn.asterisk.org/svn/asterisk/branches/13

git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@433685 65c4cc65-6c06-0410-ace0-fbb531ad65f3

4 years agoclang compiler warnings: Fix -Wunused-function; make inline function static
Matthew Jordan [Sat, 28 Mar 2015 12:20:26 +0000 (12:20 +0000)]
clang compiler warnings: Fix -Wunused-function; make inline function static

This patch fixes clang compilers warnings for unused functions. Specifically:
 * channels/chan_iax2: removed user_ref function
 * main/dsp.c: removed goertzel_update function
 * main/config.c: made variable_list_switch static

Review: https://reviewboard.asterisk.org/r/4527

ASTERISK-24917
Reported by: dkdegroot
patches:
  rb4527.patch submitted by dkdegroot (License 6600)
........

Merged revisions 433678 from http://svn.asterisk.org/svn/asterisk/branches/11
........

Merged revisions 433680 from http://svn.asterisk.org/svn/asterisk/branches/13

git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@433681 65c4cc65-6c06-0410-ace0-fbb531ad65f3

4 years agoSAC: Add conferencing extensions and configuration
Jonathan Rose [Fri, 27 Mar 2015 22:26:30 +0000 (22:26 +0000)]
SAC: Add conferencing extensions and configuration

Review: https://reviewboard.asterisk.org/r/4504/
........

Merged revisions 433656 from http://svn.asterisk.org/svn/asterisk/branches/13

git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@433657 65c4cc65-6c06-0410-ace0-fbb531ad65f3

4 years agoconfigs/basic-pbx - Super Awesome Company example configs Phase 1, Patch 2
Rusty Newton [Fri, 27 Mar 2015 21:21:05 +0000 (21:21 +0000)]
configs/basic-pbx - Super Awesome Company example configs Phase 1, Patch 2

Example configuration files for a "basic PBX" deployment for the fictitious
Super Awesome Company. Details at https://reviewboard.asterisk.org/r/4488/
and https://wiki.asterisk.org/wiki/display/AST/Super+Awesome+Company

Patch 4488 includes all functionality needed for SAC's outside connectivity
and some externally accessed features, as well as outbound dialing.

Reported by: Malcolm Davenport
Tested by: Rusty Newton

Review: https://reviewboard.asterisk.org/r/4488/
........

Merged revisions 433624 from http://svn.asterisk.org/svn/asterisk/branches/13

git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@433637 65c4cc65-6c06-0410-ace0-fbb531ad65f3

4 years agores_pjsip_registrar_expire.c: Made use ao2 container template routines and eliminated...
Richard Mudgett [Fri, 27 Mar 2015 21:06:22 +0000 (21:06 +0000)]
res_pjsip_registrar_expire.c: Made use ao2 container template routines and eliminated some RAII_VAR() usage.

* Converted the contact_autoexpire container to use the ao2 template hash
and cmp functions.  Also made use the OBJ_SEARCH_xxx names instead of the
deprecated names.

* Eliminates several unnecessary uses of RAII_VAR().

Review: https://reviewboard.asterisk.org/r/4524/
........

Merged revisions 433622 from http://svn.asterisk.org/svn/asterisk/branches/13

git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@433623 65c4cc65-6c06-0410-ace0-fbb531ad65f3

4 years agoAdd stateful PJSIP response API call, and use it for out-of-dialog responses.
Mark Michelson [Fri, 27 Mar 2015 20:46:55 +0000 (20:46 +0000)]
Add stateful PJSIP response API call, and use it for out-of-dialog responses.

Asterisk had an issue where retransmissions of MESSAGE requests resulted in
Asterisk processing the retransmission as if it were a new MESSAGE request.

This patch fixes the issue by creating a transaction in PJSIP on the incoming
request. This way, if a retransmission arrives, the PJSIP transaction layer
will resend the response and Asterisk will not ever see the retransmission.

ASTERISK-24920 #close
Reported by Mark Michelson

Review: https://reviewboard.asterisk.org/r/4532/
........

Merged revisions 433619 from http://svn.asterisk.org/svn/asterisk/branches/13

git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@433620 65c4cc65-6c06-0410-ace0-fbb531ad65f3

4 years agores_pjsip_registrar_expire.c: Cleanup scheduler leaks on unload/shutdown.
Richard Mudgett [Fri, 27 Mar 2015 20:23:58 +0000 (20:23 +0000)]
res_pjsip_registrar_expire.c: Cleanup scheduler leaks on unload/shutdown.

Contact expiration object refs were leaked when the module was unloaded.

* Made empty the scheduler of entries before destroying it to release the
object ref held by the scheduler entry.

Review: https://reviewboard.asterisk.org/r/4523/
........

Merged revisions 433596 from http://svn.asterisk.org/svn/asterisk/branches/13

git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@433617 65c4cc65-6c06-0410-ace0-fbb531ad65f3

4 years agoAdd missing file. ASTERISK-24781
Richard Mudgett [Fri, 27 Mar 2015 17:58:16 +0000 (17:58 +0000)]
Add missing file.  ASTERISK-24781

git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@433597 65c4cc65-6c06-0410-ace0-fbb531ad65f3

4 years agores/res_timing_kqueue: Update the module to conform to current timer API
Matthew Jordan [Fri, 27 Mar 2015 14:41:46 +0000 (14:41 +0000)]
res/res_timing_kqueue: Update the module to conform to current timer API

This patch updates the kqueue timing module to conform to current timer API.

This fixes issues with using the kqueue timing source on Asterisk 13 on
FreeBSD 10. These issues include:

- Remove support for kevent64().  The values used to support Asterisk timers
  fit within 32bits and so can be handled on all platforms via kevent().

- Provide debug logging for, but do not track, unacked events.  This matches
  the behavior of all other timer implementations.

- Implement continuous mode by triggering and leaving active, a user event.
  This ensures that the file descriptor for the timer returns immediately from
  poll(), without placing the load of a high speed timer on the kernel.

- In kqueue_timer_get_max_rate(), don't overstate the capability of the timer.
  On some platforms, UINT_MAX is greater than INTPTR_MAX, the largest integer
  type kqueue supports for timers.

- In kqueue_timer_get_event(), assume the caller woke up from poll() and just
  return the mode the timer is currently in. This matches all other timer
  implementations.

- Adjust the test code now that unacked events are not tracked.

Review: https://reviewboard.asterisk.org/r/4465/

ASTERISK-24857 #close
Reported by: scsiguy
Tested by: Ed Hynan
patches:
  rb4465.patch submitted by scsiguy (License 6692)
........

Merged revisions 433574 from http://svn.asterisk.org/svn/asterisk/branches/13

git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@433575 65c4cc65-6c06-0410-ace0-fbb531ad65f3

4 years agoFix link error for utils/aelparse.
Corey Farrell [Fri, 27 Mar 2015 12:27:23 +0000 (12:27 +0000)]
Fix link error for utils/aelparse.

Use the standard ast_log instead of ast_log_safe for STANDALONE programs.

Review: https://reviewboard.asterisk.org/r/4538/
........

Merged revisions 433549 from http://svn.asterisk.org/svn/asterisk/branches/11
........

Merged revisions 433550 from http://svn.asterisk.org/svn/asterisk/branches/13

git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@433551 65c4cc65-6c06-0410-ace0-fbb531ad65f3

4 years agoImproved and portable ast_log recursion avoidance
Corey Farrell [Fri, 27 Mar 2015 07:12:25 +0000 (07:12 +0000)]
Improved and portable ast_log recursion avoidance

This introduces a new logger routine ast_log_safe.  This routine should be
used for all error messages in code that can be run as a result of ast_log.
ast_log_safe does nothing if run recursively.  All error logging in
astobj2.c, strings.c and utils.h have been switched to ast_log_safe.

This required adding support for raw threadstorage.  This provides direct
access to the void* pointer in threadstorage.  In ast_log_safe, NULL is used
to signify that this thread is not already running ast_log_safe, (void*)1 when
it is already running.  This was done since it's critical that ast_log_safe
do nothing that could log during recursion checking.

ASTERISK-24155 #close
Reported by: Timo Teräs
Review: https://reviewboard.asterisk.org/r/4502/
........

Merged revisions 433522 from http://svn.asterisk.org/svn/asterisk/branches/11
........

Merged revisions 433523 from http://svn.asterisk.org/svn/asterisk/branches/13

git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@433524 65c4cc65-6c06-0410-ace0-fbb531ad65f3

4 years agoFix compile errors caused by r4500 / r4501.
Corey Farrell [Thu, 26 Mar 2015 23:09:04 +0000 (23:09 +0000)]
Fix compile errors caused by r4500 / r4501.

* Add ast_register_cleanup to utils/clicompat.c to deal with
  any utils that copy sources from main.
* Asterisk 13+: remove unused variables from core_local.c.

Review: https://reviewboard.asterisk.org/r/4534/
........

Merged revisions 433499 from http://svn.asterisk.org/svn/asterisk/branches/11
........

Merged revisions 433500 from http://svn.asterisk.org/svn/asterisk/branches/13

git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@433501 65c4cc65-6c06-0410-ace0-fbb531ad65f3

4 years agoReplace most uses of ast_register_atexit with ast_register_cleanup.
Corey Farrell [Thu, 26 Mar 2015 22:24:26 +0000 (22:24 +0000)]
Replace most uses of ast_register_atexit with ast_register_cleanup.

Since 'core stop now' and 'core restart now' do not stop modules,
it is unsafe for most of the core to run cleanups.  Originally all
cleanups used ast_register_atexit, and were only changed when it
was shown to be unsafe.  ast_register_atexit is now used only when
absolutely required to prevent corruption and close child processes.

Exceptions that need to use ast_register_atexit:
* CDR: Flush records.
* res_musiconhold: Kill external applications.
* AstDB: Close the DB.
* canary_exit: Kill canary process.

ASTERISK-24142 #close
Reported by: David Brillert

ASTERISK-24683 #close
Reported by: Peter Katzmann

ASTERISK-24805 #close
Reported by: Badalian Vyacheslav

ASTERISK-24881 #close
Reported by: Corey Farrell

Review: https://reviewboard.asterisk.org/r/4500/
Review: https://reviewboard.asterisk.org/r/4501/
........

Merged revisions 433495 from http://svn.asterisk.org/svn/asterisk/branches/11
........

Merged revisions 433497 from http://svn.asterisk.org/svn/asterisk/branches/13

git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@433498 65c4cc65-6c06-0410-ace0-fbb531ad65f3

4 years agores_pjsip: Enable unload of all modules at shutdown.
Corey Farrell [Thu, 26 Mar 2015 17:47:42 +0000 (17:47 +0000)]
res_pjsip: Enable unload of all modules at shutdown.

* Move most of res_pjsip:module_unload to unload_pjsip to resolve crashes
  caused by running PJSIP functions from non-PJSIP threads.
* Remove call to pjsip_endpt_destroy(ast_pjsip_endpoint), it was causing
  crashes in some cases.  In theory pj_shutdown() should take care of this.
* Mark res_pjsip_keepalive and res_pjsip_session as allowed to unload at
  shutdown.
* Resolve leaked config global in res_pjsip_notify.
* Unregister pubsub pjsip service module.
* Implement cleanup for res_pjsip_session.

ASTERISK-24731 #close
Reported by: Corey Farrell
Review: https://reviewboard.asterisk.org/r/4498/
........

Merged revisions 433469 from http://svn.asterisk.org/svn/asterisk/branches/13

git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@433470 65c4cc65-6c06-0410-ace0-fbb531ad65f3

4 years agoapp_confbridge: file playback blocks dtmf
Kevin Harwell [Thu, 26 Mar 2015 17:13:26 +0000 (17:13 +0000)]
app_confbridge: file playback blocks dtmf

Attempting to execute DTMF in a confbridge while file playback (prompt,
announcement, etc) is occurring is not allowed. You have to wait until
the sound file has completed before entering DTMF. This patch fixes it
so that app_confbridge now monitors for dtmf key presses during menu
driven file playback. If a key is pressed playback stops and it executes
the matched menu option.

ASTERISK-24864 #close
Reported by: Steve Pitts
Review: https://reviewboard.asterisk.org/r/4510/
........

Merged revisions 433445 from http://svn.asterisk.org/svn/asterisk/branches/11
........

Merged revisions 433446 from http://svn.asterisk.org/svn/asterisk/branches/13

git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@433447 65c4cc65-6c06-0410-ace0-fbb531ad65f3

4 years agoA couple minor cleanup tweaks.
Richard Mudgett [Wed, 25 Mar 2015 18:37:51 +0000 (18:37 +0000)]
A couple minor cleanup tweaks.

* In res/res_sorcery_realtime.c: Broke long line.

* In main/bucket.c: Eliminated unnecessary NULL check as
ast_sorcery_unref() is NULL tolerant and set the global object to NULL
after unref in the system shutdown bucket_cleanup().
........

Merged revisions 433420 from http://svn.asterisk.org/svn/asterisk/branches/13

git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@433421 65c4cc65-6c06-0410-ace0-fbb531ad65f3

4 years agores_xmpp: Buddies are always auto-registered when processing the roster
Matthew Jordan [Wed, 25 Mar 2015 15:31:06 +0000 (15:31 +0000)]
res_xmpp: Buddies are always auto-registered when processing the roster

Due to a quirk in the configuration handling of res_xmpp, the 'autoregister'
setting was never actually processed. This was due to not properly copying
over the global settings to the client settings when applying the
configuration to the run-time object.

Review: https://reviewboard.asterisk.org/r/4496/

ASTERISK-14233
ASTERISK-24780 #close
Reported by: Simon Arlott
patches:
  asterisk-13.1.0-24780 uploaded by Simon Arlott (License 5756)
........

Merged revisions 433395 from http://svn.asterisk.org/svn/asterisk/branches/11
........

Merged revisions 433396 from http://svn.asterisk.org/svn/asterisk/branches/13

git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@433397 65c4cc65-6c06-0410-ace0-fbb531ad65f3

4 years agodns: Add core DNS API + unit tests and res_resolver_unbound module + unit tests.
Joshua Colp [Wed, 25 Mar 2015 12:32:26 +0000 (12:32 +0000)]
dns: Add core DNS API + unit tests and res_resolver_unbound module + unit tests.

This change adds an abstracted core DNS API which resembles the API described
here[1]. The API provides a pluggable mechanism for resolvers and also a
consistent view for records. Both synchronous and asynchronous queries are
supported.

This change also adds a res_resolver_unbound module which uses the libunbound
library to provide resolution.

Unit tests have also been written for all of the above to confirm the API and
functionality.

ASTERISK-24834 #close
Reported by: Matt Jordan

ASTERISK-24836 #close
Reported by: Matt Jordan

Review: https://reviewboard.asterisk.org/r/4474/
Review: https://reviewboard.asterisk.org/r/4512/

[1] https://wiki.asterisk.org/wiki/display/AST/Asterisk+DNS+API

git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@433370 65c4cc65-6c06-0410-ace0-fbb531ad65f3

4 years agochan_pjsip: Add "rpid_immediate" option to prevent unnecessary "180 Ringing" messages.
Richard Mudgett [Tue, 24 Mar 2015 19:41:36 +0000 (19:41 +0000)]
chan_pjsip: Add "rpid_immediate" option to prevent unnecessary "180 Ringing" messages.

Incoming PJSIP call legs that have not been answered yet send unnecessary
"180 Ringing" or "183 Progress" messages every time a connected line
update happens.  If the outgoing channel is also PJSIP then the incoming
channel will always send a "180 Ringing" or "183 Progress" message when
the outgoing channel sends the INVITE.

Consequences of these unnecessary messages:

* The caller can start hearing ringback before the far end even gets the
call.

* Many phones tend to grab the first connected line information and refuse
to update the display if it changes.  The first information is not likely
to be correct if the call goes to an endpoint not under the control of the
first Asterisk box.

When connected line first went into Asterisk in v1.8, chan_sip received an
undocumented option "rpid_immediate" that defaults to disabled.  When
enabled, the option immediately passes connected line update information
to the caller in "180 Ringing" or "183 Progress" messages as described
above.

* Added "rpid_immediate" option to prevent unnecessary "180 Ringing" or
"183 Progress" messages.  The default is "no" to disable sending the
unnecessary messages.

ASTERISK-24781 #close
Reported by: Richard Mudgett

Review: https://reviewboard.asterisk.org/r/4473/
........

Merged revisions 433338 from http://svn.asterisk.org/svn/asterisk/branches/13

git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@433339 65c4cc65-6c06-0410-ace0-fbb531ad65f3

4 years agoFix compilations errors on 64-bit OpenBSD systems
Matthew Jordan [Mon, 23 Mar 2015 00:05:48 +0000 (00:05 +0000)]
Fix compilations errors on 64-bit OpenBSD systems

In versiong 5.5, OpenBSD went to 64-bit time values. This requires a cast to
(long) when printing members of certain time structs.

Review: https://reviewboard.asterisk.org/r/4507

ASTERISK-24879 #close
Reported by: snuffy
Tested by: snuffy
patches:
  openbsd-time64.diff uploaded by snuffy (License 5024)
........

Merged revisions 433268 from http://svn.asterisk.org/svn/asterisk/branches/11
........

Merged revisions 433269 from http://svn.asterisk.org/svn/asterisk/branches/13

git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@433270 65c4cc65-6c06-0410-ace0-fbb531ad65f3

4 years agoFix compilation issues for OpenBSD
Matthew Jordan [Sun, 22 Mar 2015 23:11:32 +0000 (23:11 +0000)]
Fix compilation issues for OpenBSD

This patch addresses compilation issues for OpenBSD. Specifically, it
addresses:
 * It allows including <sys/vmmeter.h> in asterisk.c
 * Provides a needed (size_t) cast in xmldoc.c

In 13+, it also addresses a conditional inclusion in loader.c.

Review: https://reviewboard.asterisk.org/r/4506

ASTERISK-24880 #close
Reported by: snuffy
Tested by: snuffy
patches:
  misc-openbsd.diff uploaded by snuffy (License 5024)
........

Merged revisions 433245 from http://svn.asterisk.org/svn/asterisk/branches/11
........

Merged revisions 433247 from http://svn.asterisk.org/svn/asterisk/branches/13

git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@433248 65c4cc65-6c06-0410-ace0-fbb531ad65f3

4 years agoAudit ast_pjsip_rdata_get_endpoint() usage for ref leaks.
Richard Mudgett [Fri, 20 Mar 2015 19:54:48 +0000 (19:54 +0000)]
Audit ast_pjsip_rdata_get_endpoint() usage for ref leaks.

Valgrind found some memory leaks associated with
ast_pjsip_rdata_get_endpoint().  The leaks would manifest when sending
responses to OPTIONS requests, processing MESSAGE requests, and
res_pjsip supplements implementing the incoming_request callback.

* Fix ast_pjsip_rdata_get_endpoint() endpoint ref leaks in
res/res_pjsip.c:supplement_on_rx_request(),
res/res_pjsip/pjsip_options.c:send_options_response(),
res/res_pjsip_messaging.c:rx_data_to_ast_msg(), and
res/res_pjsip_messaging.c:send_response().

* Eliminated RAII_VAR() use with ast_pjsip_rdata_get_endpoint() in
res/res_pjsip_nat.c:nat_on_rx_message().

* Fixed inconsistent but benign return value in
res/res_pjsip/pjsip_options.c:options_on_rx_request().

Review: https://reviewboard.asterisk.org/r/4511/
........

Merged revisions 433222 from http://svn.asterisk.org/svn/asterisk/branches/13

git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@433223 65c4cc65-6c06-0410-ace0-fbb531ad65f3

4 years agores_pjsip_sdp_rtp,sorcery: Fix invalid access and memory leak respectively.
Richard Mudgett [Fri, 20 Mar 2015 18:27:22 +0000 (18:27 +0000)]
res_pjsip_sdp_rtp,sorcery: Fix invalid access and memory leak respectively.

Valgrind found a memory leak and invalid access.

* Fix invalid access by sscanf() being fed a non-nul terminated string of
digits in res/res_pjsip_sdp_rtp.c:get_codecs().

* Fix memory leak in main/sorcery.c:sorcery_object_field_destructor().

* Fix potential NULL pointer dereference in
main/xmldoc.c:xmldoc_get_syntax_config_option().

Review: https://reviewboard.asterisk.org/r/4513/
........

Merged revisions 433199 from http://svn.asterisk.org/svn/asterisk/branches/13

git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@433200 65c4cc65-6c06-0410-ace0-fbb531ad65f3

4 years agofuncs/func_env: Fix regression caused in FILE read operation
Matthew Jordan [Thu, 19 Mar 2015 19:20:21 +0000 (19:20 +0000)]
funcs/func_env: Fix regression caused in FILE read operation

When r432935 was merged, it did correctly fix a situation where a FILE read
operation on the middle of a file buffer would not read the requested length
in the parameters passed to the FILE function. Unfortunately, it would also
allow the FILE function to append more bytes than what was available in the
buffer if the length exceeded the end of the buffer length.

This patch takes the minimum of the remaining bytes in the buffer along with
the calculated length to append provided by the original patch, and uses
that as the length to append in the return result. This patch also updates
the unit tests with the scenarios that were originally pointed out in
ASTERISK-21765 that the original implementation treated incorrectly.

ASTERISK-21765
........

Merged revisions 433173 from http://svn.asterisk.org/svn/asterisk/branches/11
........

Merged revisions 433174 from http://svn.asterisk.org/svn/asterisk/branches/13

git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@433175 65c4cc65-6c06-0410-ace0-fbb531ad65f3

4 years agoalemebic scripts: endpoint identifier order option
Kevin Harwell [Thu, 19 Mar 2015 15:27:56 +0000 (15:27 +0000)]
alemebic scripts: endpoint identifier order option

The script was added in 13, but when committed to trunk it caused a branch to
occur due to some trunk only alemebic changes. This fixes it so that the new
'add_pjsip_endpoint_identifier_order script points to the correct down revision.

git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@433152 65c4cc65-6c06-0410-ace0-fbb531ad65f3

4 years agologger: Apply default console logging when configuration cannot be loaded.
Corey Farrell [Thu, 19 Mar 2015 10:21:42 +0000 (10:21 +0000)]
logger: Apply default console logging when configuration cannot be loaded.

When logger.conf is missing or invalid enable console logging and display
an error message.

ASTERISK-24817 #close
Reported by: Corey Farrell
Review: https://reviewboard.asterisk.org/r/4497/
........

Merged revisions 433122 from http://svn.asterisk.org/svn/asterisk/branches/11
........

Merged revisions 433126 from http://svn.asterisk.org/svn/asterisk/branches/13

git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@433130 65c4cc65-6c06-0410-ace0-fbb531ad65f3

4 years agochan_sip: Simplify dialog/peer references, improve REF_DEBUG output.
Corey Farrell [Thu, 19 Mar 2015 09:57:59 +0000 (09:57 +0000)]
chan_sip: Simplify dialog/peer references, improve REF_DEBUG output.

* Replace functions for ref/undef of dialogs and peers with macro's
  to call ao2_t_bump/ao2_t_cleanup.
* Enable passthough of REF_DEBUG caller information to sip_alloc and
  find_call.

ASTERISK-24882 #close
Reported by: Corey Farrell
Review: https://reviewboard.asterisk.org/r/4189/
........

Merged revisions 433115 from http://svn.asterisk.org/svn/asterisk/branches/13

git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@433116 65c4cc65-6c06-0410-ace0-fbb531ad65f3

4 years agochan_sip: Fix dialog reference leaked to scheduler for reinvite_timeout.
Corey Farrell [Thu, 19 Mar 2015 09:46:06 +0000 (09:46 +0000)]
chan_sip: Fix dialog reference leaked to scheduler for reinvite_timeout.

Release the scheduler reference to the dialog for reinvite timeout during
dialog_unlink_all.

ASTERISK-24876 #close
Reported by: Corey Farrell
Review: https://reviewboard.asterisk.org/r/4491/
........

Merged revisions 433112 from http://svn.asterisk.org/svn/asterisk/branches/11
........

Merged revisions 433113 from http://svn.asterisk.org/svn/asterisk/branches/13

git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@433114 65c4cc65-6c06-0410-ace0-fbb531ad65f3

4 years agores_pjsip_session: Fix off-nominal extra unref of session.
Richard Mudgett [Wed, 18 Mar 2015 02:42:16 +0000 (02:42 +0000)]
res_pjsip_session: Fix off-nominal extra unref of session.
........

Merged revisions 433088 from http://svn.asterisk.org/svn/asterisk/branches/13

git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@433089 65c4cc65-6c06-0410-ace0-fbb531ad65f3

4 years agoVarious: bugfixes found via chaos
Scott Griepentrog [Tue, 17 Mar 2015 22:15:42 +0000 (22:15 +0000)]
Various: bugfixes found via chaos

Using DEBUG_CHAOS several instances of a null
pointer crash, and one uninitialized variable
were uncovered and fixed.  Also added details
on why Asterisk failed to initialize.

Review: https://reviewboard.asterisk.org/r/4468/
........

Merged revisions 433064 from http://svn.asterisk.org/svn/asterisk/branches/13

git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@433065 65c4cc65-6c06-0410-ace0-fbb531ad65f3

4 years agocore: Introduce chaos into memory allocations
Scott Griepentrog [Tue, 17 Mar 2015 22:03:37 +0000 (22:03 +0000)]
core: Introduce chaos into memory allocations

Locate potential crashes by exercising seldom
used code paths.  This patch introduces a new
define DEBUG_CHAOS, and mechanism to randomly
return an error condition from functions that
will seldom do so.  Functions that handle the
allocation of memory get the first treatment.

Review: https://reviewboard.asterisk.org/r/4463/
........

Merged revisions 433060 from http://svn.asterisk.org/svn/asterisk/branches/13

git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@433063 65c4cc65-6c06-0410-ace0-fbb531ad65f3

4 years agoReverting accidental ci of wrong change in r433061
Scott Griepentrog [Tue, 17 Mar 2015 22:03:01 +0000 (22:03 +0000)]
Reverting accidental ci of wrong change in r433061

git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@433062 65c4cc65-6c06-0410-ace0-fbb531ad65f3

4 years agovarious: cleanup issues found during leak hunt
Scott Griepentrog [Tue, 17 Mar 2015 22:00:05 +0000 (22:00 +0000)]
various: cleanup issues found during leak hunt

In this collection of small patches to prevent
Valgrind errors are: fixes for reference leaks
in config hooks, evaluating a parameter beyond
bounds, and accessing a structure after a lock
where it could have been already free'd.

Review: https://reviewboard.asterisk.org/r/4407/
........

Merged revisions 431583 from http://svn.asterisk.org/svn/asterisk/branches/13

git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@433061 65c4cc65-6c06-0410-ace0-fbb531ad65f3

4 years agoAudit ast_sockaddr_resolve() usage for memory leaks.
Richard Mudgett [Tue, 17 Mar 2015 21:52:47 +0000 (21:52 +0000)]
Audit ast_sockaddr_resolve() usage for memory leaks.

Valgrind found some memory leaks associated with ast_sockaddr_resolve().
Most of the leaks had already been fixed by earlier memory leak hunt
patches.  This patch performs an audit of ast_sockaddr_resolve() and found
one more.

* Fix ast_sockaddr_resolve() memory leak in
apps/app_externalivr.c:app_exec().

* Made main/netsock2.c:ast_sockaddr_resolve() always set the addrs
parameter for safety so the pointer will never be uninitialized on return.
The same goes for res/res_pjsip_acl.c:extract_contact_addr().

* Made functions that call ast_sockaddr_resolve() with RAII_VAR()
controlling the addrs variable use ast_free instead of ast_free_ptr to
provide better MALLOC_DEBUG information.

Review: https://reviewboard.asterisk.org/r/4509/
........

Merged revisions 433056 from http://svn.asterisk.org/svn/asterisk/branches/11
........

Merged revisions 433057 from http://svn.asterisk.org/svn/asterisk/branches/13

git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@433058 65c4cc65-6c06-0410-ace0-fbb531ad65f3

4 years agores_pjsip: Allow configuration of endpoint identifier query order
Kevin Harwell [Tue, 17 Mar 2015 18:35:07 +0000 (18:35 +0000)]
res_pjsip: Allow configuration of endpoint identifier query order

Updated some documentation stating that endpoint identifiers registered without
a name are place at the front of the lookup list. Also renamed register method
'ast_sip_register_endpoint_identifier_by_name' to
'ast_sip_register_endpoint_identifier_with_name'

ASTERISK-24840
Reported by: Mark Michelson
........

Merged revisions 433031 from http://svn.asterisk.org/svn/asterisk/branches/13

git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@433032 65c4cc65-6c06-0410-ace0-fbb531ad65f3

4 years agores_pjsip: Allow configuration of endpoint identifier query order
Kevin Harwell [Tue, 17 Mar 2015 18:22:20 +0000 (18:22 +0000)]
res_pjsip: Allow configuration of endpoint identifier query order

This patch fixes previously reverted code that caused binary incompatibility
problems with some modules. And like the original patch it makes sure that
no matter what order the endpoint identifier modules were loaded, priority is
given based on the ones specified in the new global 'endpoint_identifier_order'
option.

ASTERISK-24840
Reported by: Mark Michelson
Review: https://reviewboard.asterisk.org/r/4489/
........

Merged revisions 433028 from http://svn.asterisk.org/svn/asterisk/branches/13

git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@433029 65c4cc65-6c06-0410-ace0-fbb531ad65f3

4 years agores_pjsip: Add reason comment.
Richard Mudgett [Tue, 17 Mar 2015 16:11:36 +0000 (16:11 +0000)]
res_pjsip: Add reason comment.
........

Merged revisions 433005 from http://svn.asterisk.org/svn/asterisk/branches/13

git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@433006 65c4cc65-6c06-0410-ace0-fbb531ad65f3