Sean Bright [Fri, 18 Sep 2020 20:02:27 +0000 (16:02 -0400)]
res_musiconhold: Start playlist after initial announcement
Only track our sample offset if we are playing a non-announcement file,
otherwise we will skip that number of samples when we start playing the
first MoH file.
Sean Bright [Fri, 18 Sep 2020 13:09:59 +0000 (09:09 -0400)]
func_curl.c: Prevent crash when using CURLOPT(httpheader)
Because we use shared thread-local cURL instances, we need to ensure
that the state of the cURL instance is correct before each invocation.
In the case of custom headers, we were not resetting cURL's internal
HTTP header pointer which could result in a crash if subsequent
requests do not configure custom headers.
Sungtae Kim [Mon, 31 Aug 2020 12:21:09 +0000 (14:21 +0200)]
realtime: Increased reg_server character size
Currently, the ps_contacts table's reg_server column in realtime database type is varchar(20).
This is fine for normal cases, but if the hostname is longer than 20, it returns error and then
failed to register the contact address of the peer.
Normally, 20 characters limitation for the hostname is fine, but with the cloud env.
So, increased the size to 255.
Sungtae Kim [Sun, 30 Aug 2020 20:42:06 +0000 (22:42 +0200)]
res_stasis.c: Added video_single option for bridge creation
Currently, it was not possible to create bridge with video_mode single.
This made hard to put the bridge in a vidoe_single mode.
So, added video_single option for Bridge creation using the ARI.
This allows create a bridge with video_mode single.
Ben Ford [Mon, 31 Aug 2020 16:14:20 +0000 (11:14 -0500)]
Bridging: Use a ref to bridge_channel's channel to prevent crash.
There's a race condition with bridging where a bridge can be torn down
causing the bridge_channel's ast_channel to become NULL when it's still
needed. This particular case happened with attended transfers, but the
crash occurred when trying to publish a stasis message. Now, the
bridge_channel is locked, a ref to the ast_channel is obtained, and that
ref is passed down the chain.
Patrick Verzele [Tue, 1 Sep 2020 13:43:46 +0000 (14:43 +0100)]
res_pjsip_session: Deferred re-INVITE without SDP send a=sendrecv instead of a=sendonly
Building on ASTERISK-25854. When the device requests hold by sending SDP with attribute recvonly, asterisk places the session in sendonly mode. When the device later requests to resume the call by using a re-INVITE excluding SDP, asterisk needs to change the sendonly mode to sendrecv again.
Kfir Itzhak [Wed, 26 Aug 2020 09:58:21 +0000 (12:58 +0300)]
app_queue: Fix leave-empty not recording a call as abandoned
This fixes a bug introduced mistakenly in ASTERISK-25665:
If leave-empty is enabled, a call may sometimes be removed from
a queue without recording it as abandoned.
This causes Asterisk to not generate an abandon event for that
call, and for the queue abandoned counter to be incorrect.
George Joseph [Fri, 28 Aug 2020 14:34:09 +0000 (08:34 -0600)]
ast_coredumper: Fix issues with naming
If you run ast_coredumper --tarball-coredumps in the same directory
as the actual coredump, tar can fail because the link to the
actual coredump becomes recursive. The resulting tarball will
have everything _except_ the coredump (which is usually what
you need)
There's also an issue that the directory name in the tarball
is the same as the coredump so if you extract the tarball the
directory it creates will overwrite the coredump.
So:
* Made the link to the coredump use the absolute path to the
file instead of a relative one. This prevents the recursive
link and allows tar to add the coredump.
* The tarballed directory is now named <coredump>.output instead
of just <coredump> so if you expand the tarball it won't
overwrite the coredump.
Joshua C. Colp [Fri, 28 Aug 2020 09:29:23 +0000 (06:29 -0300)]
parking: Copy parker UUID as well.
When fixing issues uncovered by GCC10 a copy of the parker UUID
was removed accidentally. This change restores it so that the
subscription has the data it needs.
Tzafrir Cohen [Tue, 3 May 2016 16:11:20 +0000 (19:11 +0300)]
followme: allow disabling callee prompt
Add the option 'enable_callee_prompt' to followme.conf. Enabled by
default. If disabled, a callee is not prompted to accept or reject
the forwarded call.
Alexander Traud [Wed, 26 Aug 2020 15:43:13 +0000 (17:43 +0200)]
sip_nat_settings: Update script for latest Linux.
With the latest Linux, 'ifconfig' is not installed on default anymore.
Furthermore, the output of the current net-tools 'ifconfig' changed.
Therefore, parsing failed. This update uses 'ip addr show' instead.
Finally, the service for the external IP changed.
Alexander Traud [Wed, 26 Aug 2020 15:19:55 +0000 (17:19 +0200)]
samples: Fix keep_alive_interval default in pjsip.conf.
Since ASTERISK_27978 the default is not off but 90 seconds. That change
happened because ASTERISK_27347 disabled the keep-alives in the bundled
PJProject and Asterisk should behave the same as before.
This patch makes it so if the PJSIP_SEND_SESSION_REFRESH dialplan function
is called on a channel prior to answering a warning is issued and the
function returns unsuccessful.
Joshua C. Colp [Thu, 27 Aug 2020 10:31:40 +0000 (07:31 -0300)]
pbx: Fix hints deadlock between reload and ExtensionState.
When the ExtensionState AMI action is executed on a pattern matched
hint it can end up adding a new hint if one does not already exist.
This results in a locking order of contexts -> hints -> contexts.
If at the same time a reload is occurring and adding its own hint
it will have a locking order of hints -> contexts.
This results in a deadlock as one thread wants a lock on contexts
that the other has, and the other thread wants a lock on hints
that the other has.
This change enforces a hints -> contexts locking order by explicitly
locking hints in the places where a hint is added when queried for.
This matches the order seen through normal adding of hints.
Sean Bright [Fri, 21 Aug 2020 14:17:59 +0000 (10:17 -0400)]
app_voicemail: Process urgent messages with mailcmd
Rather than putting messages into INBOX and then moving them to Urgent
later, put them directly in to the Urgent folder. This prevents
mailcmd from being skipped.
George Joseph [Thu, 20 Aug 2020 13:57:27 +0000 (07:57 -0600)]
scope_trace: Updated macro stubs to match other branches
Although scope tracing isn't actually implemented in this branch,
macro stubs had been added to allow code that used scope tracing
to be cherry-picked to this branch without modification.
To match the other branches, the following changes were made
to the stubs...
The SCOPE_ENTER and SCOPE_EXIT* macros now print debug messages
at the same level as the scope level. This allows the same
messages to be printed to the debug log when AST_DEVMODE
isn't enabled.
Also added a few variants of the SCOPE_EXIT macros that will
also call ast_log instead of ast_debug to make it easier to
use scope tracing and still print error messages.
Sean Bright [Wed, 19 Aug 2020 17:29:51 +0000 (13:29 -0400)]
bridge_channel: Ensure text messages are zero terminated
T.140 data in RTP is not zero terminated, so when we are queuing a text
frame on a bridge we need to ensure that we are passing a zero
terminated string.
Sean Bright [Thu, 6 Aug 2020 14:58:22 +0000 (10:58 -0400)]
res_musiconhold.c: Prevent crash with realtime MoH
The MoH class internal file vector is potentially being manipulated by
multiple threads at the same time without sufficient locking. Switch to
a reference counted list and operate on copies where necessary.
Sean Bright [Tue, 4 Aug 2020 15:51:16 +0000 (11:51 -0400)]
vector.h: Fix implementation of AST_VECTOR_COMPACT() for empty vectors
The assumed behavior of realloc() - that it was effectively a free() if
its second argument was 0 - is Linux specific behavior and is not
guaranteed by either POSIX or the C specification.
Instead, if we want to resize a vector to 0, do it explicitly.
pjproject: clone sdp to protect against (nat) modifications
PJSIP, UDP transport with external_media_address and session timers
enabled. Connected to SIP server that is not in local net. Asterisk
initiated the connection and is refreshing the session after 150s
(timeout 300s). The 2nd refresh-INVITE triggered by the pjsip timer has
a malformed IP address in its SDP (garbage string). This only happens
when the SDP is modified by the nat-code to replace the local IP address
with the configured external_media_address.
Analysis: the code to modify the SDP (in
res_pjsip_session.c:session_outgoing_nat_hook() and also (redundantly?)
in res_pjsip_sdp_rtp.c:change_outgoing_sdp_stream_media_address()) uses
the tdata->pool to allocate the replacement string. But the same
pjmedia_sdp_stream that was modified for the 1st refresh-INVITE is also
used for the 2nd refresh-INVITE (because it is stored in pjmedia's
pjmedia_sdp_neg structure). The problem is, that at that moment, the
tdata->pool that holds the stringified external_media_address from the
1. refresh-INVITE has long been reused for something else.
Fix by Sauw Ming of pjproject (see
https://github.com/pjsip/pjproject/pull/2476): the local, potentially
modified pjmedia_sdp_stream is cloned in
pjproject/source/pjsip/src/pjmedia/sip_neg.c:process_answer() and the
clone is stored, thereby detaching from the tdata->pool (which is only
released *after* process_answer())
ASTERISK-28973 Reported-by: Michael Neuhauser
Change-Id: I272ac22436076596e06aa51b9fa23fd1c7734a0e
res_pjsip_registrar: Don't specify an expiration for static contacts.
Statically configured contacts on an AOR don't have an expiration
time so when adding them to the resulting 200 OK if an endpoint
registers ensure they are marked as such.
Sean Bright [Mon, 13 Jul 2020 20:06:14 +0000 (16:06 -0400)]
utf8.c: Add UTF-8 validation and utility functions
There are various places in Asterisk - specifically in regards to
database integration - where having some kind of UTF-8 validation would
be beneficial. This patch adds:
* Functions to validate that a given string contains only valid UTF-8
sequences.
* A function to copy a string (similar to ast_copy_string) stopping when
an invalid UTF-8 sequence is encountered.
* A UTF-8 validator that allows for progressive validation.
All of this is based on the excellent UTF-8 decoder by Björn Höhrmann.
More information is available here:
https://bjoern.hoehrmann.de/utf-8/decoder/dfa/
The API was written in such a way that should allow us to replace the
implementation later should we determine that we need something more
comprehensive.
pjsip: Include timer patch to prevent cancelling timer 0.
I noticed this while looking at another issue and brought
it up with Teluu. It was possible for an uninitialized timer
to be cancelled, resulting in the invalid timer id of 0
being placed into the timer heap causing issues.
This change is a backport from the pjproject repository
preventing this from happening.
res_http_websocket: Avoid reading past end of string
We read beyond the end of the buffer when copying the string out of the
buffer when we used ast_copy_string() because the original string was
not null terminated. Instead switch to ast_strndup() which does not
exhibit the same behavior.
Kevin Harwell [Tue, 30 Jun 2020 15:40:47 +0000 (10:40 -0500)]
PJSIP_MEDIA_OFFER: override what's specified on configuration
When using the PSJIP_MEDIA_OFFER dialplan function it was not
overriding an endpoint's configured codecs unless they had a
shared codec between the two.
This patch makes it so whatever is set using PJSIP_MEDIA_OFFER
is used when creating the SDP definition no matter what.
Joshua C. Colp [Fri, 26 Jun 2020 10:18:55 +0000 (07:18 -0300)]
res_pjsip: Apply AOR outbound proxy to static contacts.
The outbound proxy for an AOR was not being applied to
any statically configured Contacts. This resulted in the
OPTIONS requests being sent to the wrong target.
This change sets the outbound proxy on statically configured
contacts once the AOR configuration is done being
applied.
Joshua C. Colp [Wed, 24 Jun 2020 10:25:47 +0000 (07:25 -0300)]
menuselect: Resolve infinite loop in dependency scenario.
Given a scenario where a module has a dependency on both
an external library and a module if the external library was
available and the module was not an infinite loop would
occur. This happened due to the code changing the dependecy
status to no failure on each dependency checking loop
iteration, resulting in the code thinking that it had
gone from no failure to failure each time triggering another
dependency check.
This change makes it so that the old dependency status is
preserved throughout the dependency checking allowing it to
determine that after the first iteration the dependency
status does not transition from no failure to failure.
Frederic LE FOLL [Mon, 22 Jun 2020 09:08:47 +0000 (11:08 +0200)]
chan_sip: chan_sip does not process 400 response to an INVITE.
chan_sip handle_response() function, for a 400 response to an INVITE,
calls handle_response_invite() and does not generate ACK.
handle_response_invite() does not recognize 400 response and has no
default response processing for unexpected responses, thus it does not
generate ACK either.
The ACK on response repetition comes from handle_response() mechanism
"We must re-send ACKs to re-transmitted final responses".
According to code history, 400 response specific processing was
introduced with commit
"channels/chan_sip: Add improved support for 4xx error codes"
This commit added support for :
- 400/414/493 in handle_response_subscribe() handle_response_register()
and handle_response().
- 414/493 only in handle_response_invite().
This fix adds 400 response support in handle_response_invite().
res_corosync: Fix crash in huge distributed environment.
1) Fix memory-leaks
Added code to release ast_events extracted from corosync and stasis messages
2) Clean stasis cache when a member of the corosync cluster leaves the group
Added code to remove from the stasis cache of the members remained on the
group all the messages with the EID of the left member.
If the device states of the left member remain in the stasis cache of other
members, they will not be updated anymore and high priority cached values,
like BUSY, will take precedence over current device states.
3) Stop corosync event propagation when node is not joined to the group
Updated dispatch_thread_handler code to detect when asterisk is not joined
to the corosync group and added some condition in publish_event_to_corosync
code to send corosync messages only when joined.
When a node is not joined its corosync daemon can't send messages:
the cpg_mcast_joined function append new messages to the FIFO buffer until
it's full and then it blocks indefinitely.
In this scenario if the stasis_message_cb callback, registered by
res_corosync to handle stasis messages, try to send a corosync messages,
the thread of the stasis thread-pool will be blocked until the node join
the corosync cluster.
ASTERISK-28888
Reported by: Università di Bologna - CESIA VoIP
Guido Falsi [Thu, 18 Jun 2020 10:14:26 +0000 (12:14 +0200)]
chan_dadhi: Fix setvar in dahdi channels
The change to how setvar works for various channels performed in
ASTERISK~23756 missed some required change in the dahdi channel,
where the variables are actually set while reading configuration.
This change should fix the issue.
Joshua C. Colp [Wed, 10 Jun 2020 09:35:50 +0000 (06:35 -0300)]
res_sorcery_memory_cache: Disallow per-object expire with full backend.
The AMI action and CLI command did not take into account the properties
of full backend caching. This resulted in an expired object remaining
removed until a full backend update occurred, instead of having the
object updated when needed.
This change makes it so that the AMI action and CLI command for object
expire will now fail instead of putting the cache into an undesired
state. If full backend caching is enabled then only operations
which act on the entire cache are available.
Walter Doekes [Mon, 15 Jun 2020 11:55:37 +0000 (13:55 +0200)]
app_queue: Read latest wrapuptime instead of (possibly stale) copy
Before this changeset, it was possible that a queue member (agent) was
called even though they just got out of a call, and wrapuptime seconds
hadn't passed yet.
This could happen if a member ended a call _between_ a new call attempt
and asterisk trying that particular member for a new call.
In that case, Asterisk would check the hangup time of the
call-before-the-last-call instead of the hangup time of the-last-call.
Kevin Harwell [Fri, 15 May 2020 21:08:20 +0000 (16:08 -0500)]
pjproject: Upgrade bundled version to pjproject 2.10
This patch makes the usual necessary changes when upgrading to a new
version pjproject. For instance, version number bump, patches removed
from third-party, new *.md5 file added, etc..
This patch also includes a change to the Asterisk pjproject Makefile to
explicitly create the 'source/pjsip-apps/lib' directory. This directory
is no longer there by default so needs to be added so the Asterisk
malloc debug can be built.
This patch also includes some minor changes to Asterisk that were a result
of the upgrade. Specifically, there was a backward incompatibility change
made in 2.10 that modified the "expires header" variable field from a
signed to an unsigned value. This potentially effects comparison. Namely,
those check for a value less than zero. This patch modified a few locations
in the Asterisk code that may have been affected.
Lastly, this patch adds a new macro PJSIP_MINVERSION that can be used to
check a minimum version of pjproject at compile time.
George Joseph [Wed, 3 Jun 2020 16:23:31 +0000 (10:23 -0600)]
res_fax: Don't start a gateway if either channel is hung up
When fax_gateway_framehook is called and a gateway hasn't already
been started, the framehook gets the t38 state for both the current
channel and the peer. That call trickles down to the channel
driver which determines the state. If either channel is hung up
(or in the process of being hung up), the channel driver's tech_pvt
is going to be NULL which, in the case of chan_pjsip, will cause a
segfault.
* Added a hangup check for both the channel and peer channel
before starting a fax gateway.
* Added a check for NULL tech_pvt to chan_pjsip_queryoption
so we don't attempt to reference a tech_pvt that's already
gone.
Kevin Harwell [Wed, 3 Jun 2020 16:45:39 +0000 (11:45 -0500)]
Compiler fixes for gcc 10
This patch fixes a few compile warnings/errors that now occur when using gcc
10+.
Also, the Makefile.rules check to turn off partial inlining in gcc versions
greater or equal to 8.2.1 had a bug where it only it only checked against
versions with at least 3 numbers (ex: 8.2.1 vs 10). This patch now ensures
any version above the specified version is correctly compared.
Joshua C. Colp [Wed, 27 May 2020 08:47:07 +0000 (05:47 -0300)]
res_fax: Don't consume frames given to fax gateway on write.
In a particular fax gateway scenario whereby it would
have to translate using the read translation path on a
channel the frame being translated would be consumed.
When the frame is in the write path it is not permitted
to free the frame as the caller expects it to continue
to exist.
This change makes it so that the frame is only consumed
on the read path where it is acceptable to free it.
Pirmin Walthert [Thu, 4 Jun 2020 06:50:35 +0000 (08:50 +0200)]
res_pjsip_logger: use the correct pointer when logging tx_messages to pcap
When writing tx messages to pcap files, Asterisk is using the wrong
pointer resulting in lots of wasted space. This patch fixes it to use
the correct pointer.
George Joseph [Thu, 21 May 2020 13:20:28 +0000 (07:20 -0600)]
Scope Tracing: Add stubs to allow easier cherry-picking
Scope tracing isn't supported in Asterisk 13 due to changes made
to logging between 13 and 16 but since the scope tracing macros
may be present in the 16, 17 and master branches, those macros
are defined here as NOOPs so cherry-picking changes downward
to 13 can still be seamless.
Joshua C. Colp [Sun, 31 May 2020 22:10:29 +0000 (19:10 -0300)]
res_rtp_asterisk: Re-order RTP destruction.
The destructor for RTP deallocated transport resources
before terminating the ICE support. This could result
in a crash as the thread handling ICE would access already
freed parts of the RTP data.
This change re-orders the destruction so that ICE is
stopped before destroying things.
Pirmin Walthert [Fri, 29 May 2020 09:28:57 +0000 (11:28 +0200)]
res_pjsip_logger.c: correct the return value checks when writing to pcap
files
fwrite() does return the number of elements written and not the
number of bytes. However asterisk is currently comparing the return
value to the size of the written element what means that asterisk logs
five WARNING messages on every packet written to the pcap file.
This patch changes the code to check for the correct value, which will
always be 1.
Joshua C. Colp [Wed, 27 May 2020 14:35:42 +0000 (11:35 -0300)]
res_pjsip: Use correct pool for storing the contact_user value.
When replacing the user portion of the Contact URI the code
was using the ephemeral pool instead of the tdata pool. This
could cause the Contact user value to become invalid after a
period of time.
The code will now use the tdata pool which persists for the
lifetime of the message instead.
Pirmin Walthert [Wed, 13 May 2020 12:06:19 +0000 (14:06 +0200)]
res_pjsip_nat.c: remove x-ast-orig-host from request URI and To header
While asterisk is filtering out the x-ast-orig-host parameter from the
contact on response messages, it is not filtering it out from the
request URI and the to header on SIP requests (for example INVITE).
Joshua C. Colp [Tue, 19 May 2020 12:55:32 +0000 (09:55 -0300)]
res_sorcery_config: Always reload configuration on errors.
When a configuration file in Asterisk is loaded
information about it is stored such that on a
reload it is not reloaded if nothing has changed.
This can be problematic when an error exists in
a configuration file in PJSIP since the error
will be output at start and not subsequently on
reload if the file is unchanged.
This change makes it so that if an error is
encountered when res_sorcery_config is loading
a configuration file a reload will always read
in the configuration file, allowing the error
to be seen easier.
Joshua C. Colp [Tue, 19 May 2020 09:18:58 +0000 (06:18 -0300)]
bridge_softmix: Always remove audio from mixed frame.
When receiving audio from a channel we determine if it
is talking or silence based on a threshold value. If
this threshold is met we always mix the audio into the
conference bridge. If this threshold is not met we also
mix the audio into the conference bridge UNLESS the
drop silence option is enabled.
The code that removed the audio from the mixed frame
assumed that it was always not present if it did not
meet the threshold to be considered talking. This is
incorrect. If it has been stated that the audio was
mixed into the mixed frame then it has been mixed into
the mixed frame. By not removing audio that was
considered non-talking it was possible for a channel
to receive a slight echo of audio of itself at times.
This change ensures that the audio is always removed
from the mixed frame going back to the channel so it
no longer receives the slight echo.
res_pjsip_logger: Expand functionality to improve logging.
The PJSIP packet logger now has the following CLI commands:
pjsip set logger pcap <filename>
When used this will create a pcap file containing the incoming
and outgoing SIP packets, in unencrypted form.
pjsip set logger verbose <on / off>
This allows you to toggle logging to verbose on and off.
pjsip set logger host <IP/subnet mask> add
This allows you to add an additional IP address or subnet
mask to logging, allowing you to log multiple instead of
just a single IP address or all traffic.
The normal "pjsip set logger host" CLI command has also been
expanded to allow subnet masks as well.
Guido Falsi [Fri, 8 May 2020 11:11:47 +0000 (13:11 +0200)]
pjproject: Fix race condition when building with parallel make
Pjproject makefiles miss some dependencies which can cause race
conditions when building with parallel make processes. This patch
adds such dependencies correctly.
app.c: make sure that no non-async-signal-safe syscalls are used after
fork before exec
Posix does only allow async-signal-safe syscalls after fork before exec.
As asterisk ignores this, functions like TrySystem or System sometimes
end up in a deadlocked child process. The patch prevents the use of
non-async-signal-safe syscalls.
Guido Falsi [Sun, 3 May 2020 10:30:15 +0000 (12:30 +0200)]
pjproject: Remove bashism from configure.m4 script
The configure.m4 script for pjproject contains some += syntax, which
is specific to bash, replacing it with string substitutions makes
the script compatible with traditional Bourne shells.
ASTERISK-28866 #close Reported-by: Christoph Moench-Tegeder <cmt@FreeBSD.org>
Change-Id: I382a78160e028044598b7da83ec7e1ff42b91c05
George Joseph [Thu, 30 Apr 2020 15:56:03 +0000 (09:56 -0600)]
app_voicemail: Add workaround for a gcc 10 issue with -Wrestrict
The gcc 10 -Wrestrict option was causing "overlap" errors when
snprintf was copying one char[256] structure member to another
char[256] member in the same structure.
Using ast_alloca instead of declaring the structure inline
solves the issue.
Here's a link to the "enhancement":
https://gcc.gnu.org/legacy-ml/gcc-patches/2019-10/msg00570.html
Alexander Traud [Mon, 27 Apr 2020 15:28:01 +0000 (17:28 +0200)]
core_local: Local calls are always secure.
In a Dialplan, the channel drivers 'chan_sip' and 'chan_iax2' support
the channel items 'secure_bridge_media' and 'secure_bridge_signaling'.
That way, a channel can be forced to use encryption even if not
specified in its configuration.
However, when the Local Proxy kicks in, for example, in case of a
forwarding (SIP status 302), Local Proxy stated it does not know those
items. Consequently, such a call could not be proxied how clever your
Dialplan was. Because local calls within Asterisk are always secure,
Local Proxy accepts such a request now.
In practice it has been seen that some users come
close to our maximum ICE candidate count of 32.
In case people have gone over this increases the
count to 64, giving ample room.
res_rtp_asterisk: Protect access to nochecksums with #ifdef
Recently code accessing nochecksums variable has been added without including #ifdef SO_NO_CHECK protection, while the variable is created only when such constant is defined.
Peter Turczak [Fri, 17 Apr 2020 07:39:09 +0000 (08:39 +0100)]
chan_mobile: Add smoother to make SIP/RTP endpoints happy.
In contrast to RFC 3551, section 4.2, several SIP/RTP clients misbehave
severly (up to crashing). This patch adds another smoother for the audio
received via bt. Therefore the audio frames sent to the core will be
CHANNEL_FRAME_SIZE.
Jaco Kroon [Tue, 3 Dec 2019 18:35:20 +0000 (20:35 +0200)]
res_rtp_asterisk: iterate all local addresses looking to populate ICE.
By using pjproject to give us a list of candidates, and then filtering,
if the host has >32 addresses configured, then it is possible that we
end up filtering out all 32 of those, and ending up with no candidates
at all. Instead, get getifaddrs (which pjsip is using underlying
anyway) to retrieve all local addresses, and iterate those, adding the
first 32 addresses not excluded by the ICE ACL.
In our setup at any point in time We've got between 6 and 328 addresses
on any given system. The lower limit is the lower limit but the upper
limit is growing on a near daily basis currently.
projects / thirdparty / asterisk.git / log
