Naveen Albert [Sat, 23 Jul 2022 22:17:28 +0000 (22:17 +0000)]
pbx_functions.c: Manually update ast_str strlen.
When ast_func_read2 is used to read a function using
its read function (as opposed to a native ast_str read2
function), the result is copied directly by the function
into the ast_str buffer. As a result, the ast_str length
remains initialized to 0, which is a bug because this is
not the real string length.
This can cascade and have issues elsewhere, such as when
reading substrings of functions that only register read
as opposed to read2 callbacks. In this case, since reading
ast_str_strlen returns 0, the returned substring is empty
as opposed to the actual substring. This has caused
the ast_str family of functions to behave inconsistently
and erroneously, in contrast to the pbx_variables substitution
functions which work correctly.
This fixes this issue by manually updating the ast_str length
when the result is copied directly into the ast_str buffer.
Additionally, an assertion and a unit test that previously
exposed these issues are added, now that the issue is fixed.
build: fix bininstall launchd issue on cross-platform build
configure script detects /sbin/launchd, but the result of this
check is not used in Makefile (bininstall). Makefile also detects
/sbin/launchd file to decide if it is required to install
safe_asterisk.
configure script correctly detects cross compile build and sets
PBX_LAUNCHD=0
In case of building asterisk on MacOS host for Linux target using
external toolchain (e.g. OpenWrt toolchain), bininstall does not
install safe_asterisk (due to /sbin/launchd detection in Makefile),
but it is required on target (Linux).
This patch adds HAVE_SBIN_LAUNCHD=@PBX_LAUNCHD@ to makeopts.in to
use the result of /sbin/launchd detection from configure script in
Makefile.
Also this patch uses HAVE_SBIN_LAUNCHD in Makefile (bininstall) to
decide if it is required to install safe_asterisk.
Naveen Albert [Tue, 12 Jul 2022 21:38:12 +0000 (21:38 +0000)]
manager: Fix incomplete filtering of AMI events.
The global event filtering code was only in one
possible execution path, so not all events were
being properly filtered out if requested. This moves
that into the universal AMI handling code so all
events are properly handled.
Additionally, the CLI listing of disabled events can
also get truncated, so we now print out everything.
Naveen Albert [Mon, 11 Jul 2022 11:32:39 +0000 (11:32 +0000)]
db: Add AMI action to retrieve DB keys at prefix.
Adds the DBGetTree action, which can be used to
retrieve all of the DB keys beginning with a
particular prefix, similar to the capability
provided by the database show CLI command.
res_pjsip: delay contact pruning on Asterisk start
Move the call to ast_sip_location_prune_boot_contacts() *after* the call
to ast_res_pjsip_init_options_handling() so that
res/res_pjsip/pjsip_options.c is informed about the contact deletion and
updates its sip_options_contact_statuses list. This allows for an AMI
event to be sent by res/res_pjsip/pjsip_options.c if the endpoint
registers again from the same remote address and port (i.e., same URI)
as used before the Asterisk restart.
ASTERISK-30109 Reported-by: Michael Neuhauser
Change-Id: I1ba4478019e4931a7085f62708d9b66837e901a8
Naveen Albert [Tue, 29 Mar 2022 01:35:43 +0000 (01:35 +0000)]
chan_dahdi: Fix buggy and missing Caller ID parameters
There are several things wrong with analog Caller ID
handling that are fixed by this commit:
callerid.c's Caller ID generation function contains the
logic to use the presentation to properly send the proper
Caller ID. However, currently, DAHDI does not pass any
presentation information to the Caller ID module, which
means that presentation is completely ignored on all calls.
This means that lines could be getting Caller ID information
they aren't supposed to.
Part of the reason this has been obscured is because the
simple switch logic for handling the built in *67 and *82
is completely wrong. Rather than modifying the presentation
for the call accordingly (which is what it's supposed to do),
it simply blanks out the Caller ID or fills it in. This is
wrong, so wrong that it makes a mockery of the specification.
Additionally, it would leave to the "UNAVAILABLE" disposition
being used for Caller ID generation as opposed to the "PRIVATE"
disposition that it should have been using. This is now fixed
to only update the presentation and not modify the number and
name, so that the simple switch *67/*82 work correctly.
Next, sig_analog currently only copies over the name and number,
nothing else, when it is filling in a duplicated caller id
structure. Thus, we also now copy over the presentation
information so that is available for the Caller ID spill.
Additionally, this meant that "valid" was implicitly 0,
and as such presentation would always fail to "Unavailable".
The validity is therefore also copied over so it can be used
by ast_party_id_presentation.
As part of this fix, new API is added so that all the relevant
Caller ID information can be passed in to the Caller ID generation
functions. Parameters that are also completely missing from the
Caller ID spill have also been added, to enhance the compatibility,
correctness, and completeness of the Asterisk Caller ID implementation.
Naveen Albert [Fri, 1 Apr 2022 20:17:15 +0000 (20:17 +0000)]
chan_dahdi: Add POLARITY function.
Adds a POLARITY function which can be used to
retrieve the current polarity of an FXS channel
as well as set the polarity of an FXS channel
to idle or reverse at any point during a call.
Mike Bradeen [Thu, 2 Jun 2022 02:03:06 +0000 (20:03 -0600)]
Makefile: Avoid git-make user conflict
make_version now silently checks if the required git commands will
fail. If they do, then return UNKNOWN__git_check_fail to
distinguish this failure from other UNKNOWN__ version failures
Makefile checks for this value on install and exits out with
instructions
Naveen Albert [Sat, 18 Jun 2022 12:17:50 +0000 (12:17 +0000)]
app_confbridge: Always set minimum video update interval.
Currently, if multiple video-enabled ConfBridges are
conferenced together, we immediately get into a scenario
where an infinite sequence of video updates fills up
the taskprocessor queue and causes memory consumption
to climb unabated until Asterisk is killed. This is due
to the core bridging mechanism that provides video updates
(softmix_bridge_write_control in bridge_softmix.c)
continously updating all the channels in the bridge with
video updates.
The logic to do so in the core is that the video updates
should be provided if the video_update_discard property
for the bridge is 0, or if enough time has elapsed since
the last video update. Thus, we already have a safeguard
built in to ensure the scenario described above does not
happen. Currently, however, this safeguard is not being
adequately ensured.
In app_confbridge, the video_update_discard property
defaults to 2000, which is a healthy value that should
completely prevent this issue. However, this value is
only set onto the bridge in the SFU video mode. This
leaves video modes such as follow_talker completely
vulnerable, since video_update_discard will actually
be 0, since the default or set value was never applied.
As a result, the core bridging mechanism will always
try to provide video updates regardless of when the last
one was sent.
To prevent this issue from happening, we now always
set the video_update_discard property on the bridge
with the value from the bridge profile. The app_confbridge
defaults will thus ensure that infinite video updates
no longer happen in any video mode.
Sean Bright [Tue, 5 Jul 2022 15:24:58 +0000 (11:24 -0400)]
pbx.c: Simplify ast_context memory management.
Allocate all of the ast_context's character data in the structure's
flexible array member and eliminate the clunky fake_context. This will
simplify future changes to ast_context.
George Joseph [Wed, 13 Jul 2022 18:38:10 +0000 (12:38 -0600)]
geoloc_eprofile.c: Fix setting of loc_src in set_loc_src()
line 196: loc_src = '\0';
should have been
line 196: *loc_src = '\0';
The issue was caught by the gcc optimizer complaining that
loc_src had a zero length because the pointer itself was being
set to NULL instead of the _contents_ of the pointer being set
to the NULL terminator.
ASTERISK-30138 Reported-by: Sean Bright
Change-Id: Id247be113cc8510f043ca053d5b4f5f3d32acd29
George Joseph [Thu, 7 Jul 2022 15:32:38 +0000 (09:32 -0600)]
Geolocation: chan_pjsip Capability Preview
This commit adds res_pjsip_geolocation which gives chan_pjsip
the ability to use the core geolocation capabilities.
This commit message is intentionally short because this isn't
a simple capability. See the documentation at
https://wiki.asterisk.org/wiki/display/AST/Geolocation
for more information.
THE CAPABILITIES IMPLEMENTED HERE MAY CHANGE BASED ON
USER FEEDBACK!
George Joseph [Tue, 15 Feb 2022 13:29:50 +0000 (06:29 -0700)]
Geolocation: Core Capability Preview
This commit adds res_geolocation which creates the core capabilities
to manipulate Geolocation information on SIP INVITEs.
An upcoming commit will add res_pjsip_geolocation which will
allow the capabilities to be used with the pjsip channel driver.
This commit message is intentionally short because this isn't
a simple capability. See the documentation at
https://wiki.asterisk.org/wiki/display/AST/Geolocation
for more information.
THE CAPABILITIES IMPLEMENTED HERE MAY CHANGE BASED ON
USER FEEDBACK!
Naveen Albert [Mon, 4 Jul 2022 10:21:24 +0000 (10:21 +0000)]
chan_iax2: Allow compiling without OpenSSL.
ASTERISK_30007 accidentally made OpenSSL a
required depdendency. This adds an ifdef so
the relevant code is compiled only if OpenSSL
is available, since it only needs to be executed
if OpenSSL is available anyways.
A sporadic test failure was happening when executing the AEAP
Websocket transport tests. It was originally thought this was
due to things not getting cleaned up fast enough, but upon further
investigation I determined the underlying cause was poll()
getting interrupted and this not being handled in all places.
This change adds EINTR and EAGAIN handling to the Websocket
client connect code as well as the AEAP Websocket transport code.
If either occur then the code will just go back to waiting
for data.
The originally disabled failure test case has also been
re-enabled.
Naveen Albert [Sat, 14 May 2022 21:25:04 +0000 (21:25 +0000)]
res_cliexec: Add dialplan exec CLI command.
Adds a CLI command similar to "dialplan eval function" except for
applications: "dialplan exec application", useful for quickly
testing certain application behavior directly from the CLI
without writing any dialplan.
features: Update documentation for automon and automixmon
The current documentation is out of date and does not reflect actual
behaviour. This change makes documentation clearer and accurately
reflect the purpose of relevant channel variables.
Naveen Albert [Sat, 30 Apr 2022 16:44:26 +0000 (16:44 +0000)]
app_dial: Fix dial status regression.
ASTERISK_28638 caused a regression by incorrectly aborting
early and overwriting the status on certain calls.
This was exhibited by certain technologies such as DAHDI,
where DAHDI returns NULL for the request if a line is busy.
This caused the BUSY condition to be incorrectly treated
as CHANUNAVAIL because the DIALSTATUS was getting incorrectly
overwritten and call handling was aborted early.
This is fixed by instead checking if any valid peers have been
specified, as opposed to checking the list size of successful
requests. This is because the latter could be empty but this
does not indicate any kind of problem. This restores the
previous working behavior.
Naveen Albert [Fri, 1 Apr 2022 19:49:31 +0000 (19:49 +0000)]
db: Notify user if deleted DB entry didn't exist.
Currently, if using the CLI to delete a DB entry,
"Database entry removed" is always returned,
regardless of whether or not the entry actually
existed in the first place. This meant that users
were never told if entries did not exist.
The same issue occurs if trying to delete a DB key
using AMI.
To address this, new API is added that is more stringent
in deleting values from AstDB, which will not return
success if the value did not exist in the first place,
and will print out specific error details if available.
Naveen Albert [Sat, 18 Jun 2022 17:13:07 +0000 (17:13 +0000)]
app_dial: Propagate outbound hook flashes.
The Dial application currently stops hook flashes
dead in their tracks from propagating through on
outbound calls. This fixes that so they can go
down the wire.
Naveen Albert [Sat, 5 Feb 2022 21:16:27 +0000 (21:16 +0000)]
cli: Fix CLI blocking forever on terminating backslash
A corner case exists in CLI parsing where if
a CLI user in a remote console ends with
a backslash and then invokes command completion
(using TAB or ?), then the console will freeze
forever until a SIGQUIT signal is sent to the
process, due to getting blocked forever
reading the command completion. CTRL+C
and other key combinations have no impact on
the CLI session.
This occurs because, in such cases, the CLI
process is waiting for AST_CLI_COMPLETE_EOF
to appear in the buffer from the main process,
but instead the main process is confused by
the funny syntax and thus prints out the CLI help.
As a result, the CLI process is stuck on the
read call, waiting for the completion that
will never come.
This prevents blocking forever by checking
if the data from the main process starts with
"Usage:". If it does, that means that CLI help
was sent instead of the tab complete vector,
and thus the CLI should bail out and not wait
any longer.
Kevin Harwell [Wed, 8 Jun 2022 23:32:33 +0000 (18:32 -0500)]
res_pjsip: allow TLS verification of wildcard cert-bearing servers
Rightly the use of wildcards in certificates is disallowed in accordance
with RFC5922. However, RFC2818 does make some allowances with regards to
their use when using subject alt names with DNS name types.
As such this patch creates a new setting for TLS transports called
'allow_wildcard_certs', which when it and 'verify_server' are both enabled
allows DNS name types, as well as the common name that start with '*.'
to match as a wildcard.
For instance: *.example.com
will match for: foo.example.com
Partial matching is not allowed, e.g. f*.example.com, foo.*.com, etc...
And the starting wildcard only matches for a single level.
For instance: *.example.com
will NOT match for: foo.bar.example.com
Naveen Albert [Sun, 15 May 2022 12:41:06 +0000 (12:41 +0000)]
pbx: Add helper function to execute applications.
Finding an application and executing it if found is
a common task throughout Asterisk. This adds a helper
function around pbx_exec to do this, to eliminate
redundant code and make it easier for modules to
substitute variables and execute applications by name.
Naveen Albert [Sat, 11 Jun 2022 19:29:55 +0000 (19:29 +0000)]
asterisk.c: Fix incompatibility warnings for remote console.
A previous review fixing ASTERISK_22246 and ASTERISK_26582
got a couple of the options mixed up as to whether or not
they are compatible with the remote console. This fixes
those to the best of my knowledge.
Kevin Harwell [Tue, 7 Jun 2022 21:03:21 +0000 (16:03 -0500)]
test_aeap_transport: disable part of failing unit test
The 'transport_binary' test sporadically fails, but on a theory that the
problem is caused by a previously executed test, transport_connect_fail,
part of that test has been disabled until a solution is found.
Naveen Albert [Fri, 13 May 2022 12:33:21 +0000 (12:33 +0000)]
sig_analog: Fix broken three-way conferencing.
Three-way calling for analog lines is currently broken.
If party A is on a call with party B and initiates a
three-way call to party C, the behavior differs depending
on whether the call is conferenced prior to party C
answering. The post-answer case is correct. However,
if A flashes before C answers, then the next flash
disconnects B rather than C, which is incorrect.
This error occurs because the subs are not swapped
in the misbehaving case. This is because the flash
handler only swaps the subs if C has answered already,
which is wrong. To fix this, we swap the subs regardless
of whether C has answered or not when the call is
conferenced. This ensures that C is disconnected
on the next hook flash, rather than B as can happen
currently.
Naveen Albert [Sun, 15 May 2022 13:31:09 +0000 (13:31 +0000)]
app_voicemail: Add option to prevent message deletion.
Adds an option to VoiceMailMain that prevents the user
from deleting messages during that application invocation.
This can be useful for public or shared mailboxes, where
some users should be able to listen to messages but not
delete them.
Trevor Peirce [Tue, 7 Jun 2022 21:53:34 +0000 (14:53 -0700)]
res_pjsip: Actually enable session timers when timers=always
When a pjsip endpoint is defined with timers=always, this has been a
functional noop. This patch correctly sets the feature bitmap to both
enable support for session timers and to enable them even when the
endpoint itself does not request or support timers.
ASTERISK-29603 Reported-By: Ray Crumrine
Change-Id: I8b5eeaa9ec7f50cc6d96dd34c2b4aa9c53fb5440
app_sayunixtime: Use correct inflection for German time.
In function ast_say_date_with_format_de(), take special
care when the hour is one o'clock. In this case, the
German number "eins" must be inflected to its neutrum form,
"ein". This is achieved by playing "digits/1N" instead of
"digits/1". Fixes both 12- and 24-hour formats.
Naveen Albert [Sat, 12 Mar 2022 18:27:41 +0000 (18:27 +0000)]
res_pjsip_outbound_registration: Make max random delay configurable.
Currently, PJSIP will randomly wait up to 10 seconds for each
outbound registration's initial attempt. The reason for this
is to avoid having all outbound registrations attempt to register
simultaneously.
This can create limitations with the test suite where we need to
be able to receive inbound calls potentially within 10 seconds of
starting up. For instance, we might register to another server
and then try to receive a call through the registration, but if
the registration hasn't happened yet, this will fail, and hence
this inconsistent behavior can cause tests to fail. Ultimately,
this requires a smaller random value because there may be no good
reason to wait for up to 10 seconds in these circumstances.
To address this, a new config option is introduced which makes this
maximum delay configurable. This allows, for instance, this to be
set to a very small value in test systems to ensure that registrations
happen immediately without an unnecessary delay, and can be used more
generally to control how "tight" the initial outbound registrations
are.
Naveen Albert [Mon, 16 May 2022 13:01:48 +0000 (13:01 +0000)]
chan_iax2: Prevent deadlock due to duplicate autoservice.
If a switch is invoked using chan_iax2, deadlock can result
because the PBX core is autoservicing the channel while chan_iax2
also then attempts to service it while waiting for the result
of the switch. This removes servicing of the channel to prevent
any conflicts.
Naveen Albert [Wed, 23 Mar 2022 11:05:28 +0000 (11:05 +0000)]
res_calendar: Prevent assertion if event ends in past.
res_calendar will trigger an assertion currently
if the ending time is calculated to be in the past.
Unlike the reminder and start times, however, there
is currently no check to catch non-positive times
and set them to 1. As a result, if we get a negative
value by happenstance, this can cause a crash.
To prevent the assertion from begin triggered, we now
use the same logic as the reminder and start events
to catch this issue before it can cause a problem.
Naveen Albert [Tue, 3 May 2022 12:44:07 +0000 (12:44 +0000)]
loader: Prevent deadlock using tab completion.
If tab completion using ast_module_helper is attempted
during startup, deadlock will ensue because the CLI
will attempt to lock the module list while it is already
locked by the loader. This causes deadlock because when
the loader tries to acquire the CLI lock, they are blocked
on each other.
Waiting for startup to complete is not feasible because
the CLI lock is acquired while waiting, so deadlock will
ensure regardless of whether or not a lock on the module
list is attempted.
To prevent deadlock, we immediately abort if tab completion
is attempted on the module list before Asterisk is fully
booted.
chan_pjsip: Only set default audio stream on hold.
When a PJSIP channel is set on hold or off hold, all streams were set
on/off hold. This is not the desired behaviour and caused issues
when there were multiple streams in the topology.
Now, only the default audio stream is set on/off hold when a hold is
indicated.
Alexei Gradinari [Thu, 26 May 2022 21:29:29 +0000 (17:29 -0400)]
res_pjsip_dialog_info_body_generator: Set LOCAL target URI as local URI
The change "Add LOCAL/REMOTE tags in dialog-info+xml" set both "local"
Identity Element URI and Target Element URI to the same value -
the channel Caller Number.
For Identity Element it's ok to set as Caller ID.
But Local Target URI should be set as local URI.
In this case the Local Target URI can be used for Directed Call Pickup
by Polycom ip-phones (parameter useLocalTargetUriforLegacyPickup).
res_agi: Evaluate dialplan functions and variables in agi exec if enabled
Agi commnad exec can now evaluate dialplan functions and
variables if variable AGIEXECFULL is set to yes. this can
be useful when executing Playback or Read from agi.
Moritz Fain [Mon, 25 Apr 2022 22:40:49 +0000 (00:40 +0200)]
ari: expose channel driver's unique id to ARI channel resource
This change exposes the channel driver's unique id (i.e. the Call-ID
for chan_sip/chan_pjsip based channels) to ARI channel resources
as `protocol_id`.
ASTERISK-30027
Reported by: Moritz Fain
Tested by: Moritz Fain
Joshua C. Colp [Tue, 17 May 2022 12:18:03 +0000 (09:18 -0300)]
res_pjsip_transport_websocket: Also set the remote name.
As part of PJSIP 2.11 a behavior change was done to require
a matching remote hostname on an established transport for
secure transports. Since the Websocket transport is considered
a secure transport this caused the existing connection to not
be found and used.
We now set the remote hostname and the transport can be found.
Naveen Albert [Fri, 29 Apr 2022 16:42:55 +0000 (16:42 +0000)]
app_confbridge: Add function to retrieve channels.
Adds the CONFBRIDGE_CHANNELS function which can be used
to retrieve a comma-separated list of channels, filtered
by a particular type of participant category. This output
can then be used with functions like UNSHIFT, SHIFT, POP,
etc.
Naveen Albert [Tue, 26 Apr 2022 19:00:14 +0000 (19:00 +0000)]
chan_dahdi: Fix broken operator mode clearing.
Currently, the operator services mode in DAHDI is broken and unusable.
The actual operator recall functionality works properly; however,
when the operator hangs up (which is the only way that such a call
is allowed to end), both lines are permanently taken out of service
until "dahdi restart" is run. This prevents this feature from being
used.
Operator mode is one of the few factors that can cause the general
analog event handling in sig_analog not to be used. Several years
back, much of the analog handling was moved from chan_dahdi to
sig_analog. However, this was not done fully or consistently at
the time, and when operator mode is active, sig_analog does not
get used. Generally this is correct, but in the case of hangup
it should be using sig_analog regardless of the operator mode;
otherwise, the lines do not properly clear and they become unusable.
This bug is fixed so the operator can now hang up and properly
release the call. It is treated just like any other hangup. The
operator mode functionality continues to work as it did before.
George Joseph [Tue, 3 May 2022 12:57:58 +0000 (06:57 -0600)]
GCC12: Fixes for 16+
Most issues were in stringfields and had to do with comparing
a pointer to an constant/interned string with NULL. Since the
string was a constant, a pointer to it could never be NULL so
the comparison was always "true". gcc now complains about that.
There were also a few issues where determining if there was
enough space for a memcpy or s(n)printf which were fixed
by defining some of the involved variables as "volatile".
core_unreal: Flip stream direction of second channel.
When a new unreal (local) channel is created, a second (;2) channel is
created as a counterpart which clones the topology of the first
channel. This creates issues when an outgoing stream is sendonly or
recvonly as the stream state of the inbound channel will be the same
as the stream state of the outbound channel.
Now the stream state is flipped for the streams of the 2nd channel in
ast_unreal_new_channels if the outgoing stream topology is recvonly or
sendonly.
Naveen Albert [Sun, 27 Mar 2022 12:33:39 +0000 (12:33 +0000)]
chan_dahdi: Document dial resource options.
Documents the Dial syntax for DAHDI, namely the channel group,
distinctive ring, answer confirmation, and digital call options
that are specified in the resource itself.
Naveen Albert [Tue, 29 Mar 2022 23:47:41 +0000 (23:47 +0000)]
chan_dahdi: Don't allow MWI FSK if channel not idle.
For lines that have mailboxes configured on them, with
FSK MWI, DAHDI will periodically try to dispatch FSK
to update MWI. However, this is never supposed to be
done when a channel is not idle.
There is currently an edge case where MWI FSK can
extraneously get spooled for the channel if a caller
hook flashes and hangs up, which triggers a recall ring.
After one ring, the on hook time threshold in this if
condition has been satisfied and an MWI update is spooled.
This means that when the phone is picked up again, the
answerer gets an FSK spill before being reconnected to
the party on hold.
To prevent this, we now explicitly check to ensure that
subchannel 0 has no owner. There is no owner when DAHDI
channels are idle, but if the channel is "in use" in some
way (such as in the aforementioned scenario), then there
is an owner, and we shouldn't process MWI at this time.
Michael Cargile [Wed, 23 Feb 2022 16:29:09 +0000 (11:29 -0500)]
apps/confbridge: Added hear_own_join_sound option to control who hears sound_join
Added the hear_own_join_sound option to the confbridge user profile to
control who hears the sound_join audio file. When set to 'yes' the user
entering the conference and the participants already in the conference
will hear the sound_join audio file. When set to 'no' the user entering
the conference will not hear the sound_join audio file, but the
participants already in the conference will hear the sound_join audio
file.
Naveen Albert [Sun, 27 Mar 2022 11:23:12 +0000 (11:23 +0000)]
chan_dahdi: Don't append cadences on dahdi restart.
Currently, if any custom ring cadences are specified, they are
appended to the array of cadences from wherever we left off
last time. This works properly the first time, but on subsequent
dahdi restarts, it means that the existing cadences are left
alone and (most likely) the same cadences are then re-added
afterwards. In short order, the cadence array gets maxed out
and the user begins seeing warnings that the array is full
and no more cadences may be added.
This buggy behavior persists until Asterisk is completely
restarted; however, if and when dahdi restart is run again,
then the same problem is reintroduced.
This fixes this behavior so that cadence parsing is more
idempotent, that is so running dahdi restart multiple times
starts adding cadences from the beginning, rather than from
wherever the last cadence was added.
As before, it is still not possible to revert to the default
cadences by simply removing all cadences in this manner, nor
is it possible to delete existing cadences. However, this
does make it possible to update existing cadences, which
was not possible before, and also ensures that the cadences
remain unchanged if the config remains unchanged.
Naveen Albert [Sat, 2 Apr 2022 21:22:14 +0000 (21:22 +0000)]
chan_iax2: Prevent crash if dialing RSA-only call without outkey.
Currently, if attempting to place a call to a peer that only allows
RSA authentication, if we fail to provide an outkey when placing
the call, Asterisk will crash.
This exposes the broader issue that IAX2 is prone to causing a crash
if encryption or decryption is attempted but we never initialized
the encryption and decryption keys. In other words, if the logic
to use encryption in chan_iax2 is not perfectly aligned with the
decision to build keys in the first place, then a crash is not
only possible but probable. This was demonstrated by ASTERISK_29264,
for instance.
This permanently prevents such events from causing a crash by explicitly
checking that keys are initialized properly before setting the flags
to use encryption for the call. Instead of crashing, the call will
now abort.
Naveen Albert [Sat, 5 Feb 2022 15:03:21 +0000 (15:03 +0000)]
menuselect: Don't erroneously recompile modules.
A bug in menuselect can cause modules that are disabled
by default to be recompiled every time a recompilation
occurs. This occurs for module categories that are NOT
positive output, as for these categories, the modules
contained in the makeopts file indicate modules which
should NOT be selected. The existing procedure of iterating
through these modules to mark modules as present is thus
insufficient. This has led to modules with a default_enabled
tag of "no" to get deleted and recompiled every time, even
when they haven't changed.
To fix this, we now modify the mark as present behavior
for module categories that are not positive output. For
these, we start by iterating through the module tree
and marking all modules as present, then go back and
mark anything contained in the makeopts file as not
present. This ensures that makeopt selections are actually
used properly, regardless of whether a module category
uses positive output or not.
Naveen Albert [Thu, 31 Mar 2022 15:44:23 +0000 (15:44 +0000)]
app_meetme: Don't erroneously set global variables.
The admin_exec function in app_meetme is used by the SLA
applications for internal bridging. However, in these cases,
chan is NULL. Currently, this function will set some status
variables that are intended for a channel, but since channel
is NULL, this is erroneously creating meaningless global
variables, which shouldn't be happening. This sets these
variables only if chan is not NULL.
Naveen Albert [Wed, 30 Mar 2022 00:22:46 +0000 (00:22 +0000)]
chan_dahdi: Fix insufficient array size for round robin.
According to chan_dahdi.conf, up to 64 groups (numbered
0 through 63) can be used when dialing DAHDI channels.
However, currently dialing round robin with a group number
greater than 31 fails because the array for the round robin
structure is only size 32, instead of 64 as it should be.
This fixes that so the round robin array size is consistent
with the actual groups capacity.
Naveen Albert [Sat, 5 Mar 2022 11:43:13 +0000 (11:43 +0000)]
asterisk.c: Warn of incompatibilities with remote console.
Some command line options to Asterisk only apply when Asterisk
is started and cannot be used with remote console mode. If a
user tries to use any of these, they are currently simply
silently ignored.
This prints out a warning if incompatible options are used,
informing users that an option used cannot be used with remote
console mode. Additionally, some clarifications are added to
the help text and man page.
Naveen Albert [Tue, 1 Mar 2022 01:29:43 +0000 (01:29 +0000)]
file.c: Prevent formats from seeking negative offsets.
Currently, if a user uses an application like ControlPlayback
to try to rewind a file past the beginning, this can throw
warnings when the file format (e.g. PCM) tries to seek to
a negative offset.
Instead of letting file formats try (and fail) to seek a
negative offset, we instead now catch this in the rewind
function to ensure that we never seek an offset less than 0.
This prevents legitimate user actions from triggering warnings
from any particular file formats.
Naveen Albert [Sat, 26 Feb 2022 12:37:08 +0000 (12:37 +0000)]
chan_pjsip: Add ability to send flash events.
PJSIP currently is capable of receiving flash events
and converting them to FLASH control frames, but it
currently lacks support for doing the reverse: taking
a FLASH control frame and converting it into a flash
event in the SIP domain.
This adds the ability for PJSIP to process flash control
frames by converting them into the appropriate SIP INFO
message, which can then be sent to the peer. This allows,
for example, flash events to be sent between Asterisk
systems using PJSIP.
Naveen Albert [Sun, 26 Dec 2021 21:39:12 +0000 (21:39 +0000)]
cli: Add command to evaluate dialplan functions.
Adds the dialplan eval function commands to evaluate a dialplan
function from the CLI. The return value and function result are
printed out and can be used for testing or debugging.
On a write error to an AMI session a flag was set to
indicate that the write error had occurred, with the
expected result being that the session be terminated.
This was not actually happening and instead writing
would continue to be attempted.
This change adds a check for the write error and causes
the session to actually terminate.
Add framework to connect to, and read and write protocol based
messages from and to an external application using an Asterisk
External Application Protocol (AEAP). This has been divided into
several abstractions:
1. transport - base communication layer (currently websocket only)
2. message - AEAP description and data (currently JSON only)
3. transaction - links/binds requests and responses
4. aeap - transport, message, and transaction handler/manager
This patch also adds an AEAP implementation for speech to text.
Existing speech API callbacks for speech to text have been completed
making it possible for Asterisk to connect to a configured external
translator service and provide audio for STT. Results can also be
received from the external translator, and made available as speech
results in Asterisk.
Unit tests have also been created that test the AEAP framework, and
also the speech to text implementation.
Ben Ford [Mon, 29 Mar 2021 17:28:24 +0000 (12:28 -0500)]
res_aeap: Add basic config skeleton and CLI commands.
Added support for a basic AEAP configuration read from aeap.conf.
Also added 2 CLI commands for showing individual configurations as
well as all of them: aeap show server <id> and aeap show servers.
Only one configuration option is required at the moment, and that one is
server_url. It must be a websocket URL. The other option, codecs, is
optional and will be used over the codecs specified on the endpoint if
provided.
The async_operations setting on a transport configures how
many simultaneous incoming packets the transport can handle
when multiple threads are polling and waiting on the transport.
As we only use a single thread this was needlessly creating
incoming packets when set to a non-default value, wasting memory.
Ben Ford [Thu, 21 Apr 2022 15:26:01 +0000 (10:26 -0500)]
res_pjsip_stir_shaken.c: Fix enabled when not configured.
There was an issue with the conditional where STIR/SHAKEN would be
enabled even when not configured. It has been changed to ensure that if
a profile does not exist and stir_shaken is not set in pjsip.conf, then
the conditional will return from the function without performing
STIR/SHAKEN operations.
app_dial: Flip stream direction of outgoing channel.
When executing dial, the topology of the incoming channel is cloned and
used for the outgoing channel. This creates issues when an incoming
stream is sendonly or recvonly as the stream state of the outgoing
channel will be the same as the stream state of the incoming channel.
Now the stream state is flipped for the outgoing stream in
dial_exec_full if the incoming stream topology is recvonly or sendonly.
Ben Ford [Mon, 28 Feb 2022 17:19:54 +0000 (11:19 -0600)]
AST-2022-002 - res_stir_shaken/curl: Add ACL checks for Identity header.
Adds a new configuration option, stir_shaken_profile, in pjsip.conf that
can be specified on a per endpoint basis. This option will reference a
stir_shaken_profile that can be configured in stir_shaken.conf. The type
of this option must be 'profile'. The stir_shaken option can be
specified on this object with the same values as before (attest, verify,
on), but it cannot be off since having the profile itself implies wanting
STIR/SHAKEN support. You can also specify an ACL from acl.conf (along
with permit and deny lines in the object itself) that will be used to
limit what interfaces Asterisk will attempt to retrieve information from
when reading the Identity header.
projects / thirdparty / asterisk.git / log
