Matthew Jordan [Mon, 8 Jul 2013 14:26:40 +0000 (14:26 +0000)]
Create Local channel messages on the Stasis message bus and produce AMI events
This patch does the following:
* It adds a virtual table of callbacks to core_unreal. These callbacks can be
supplied by concrete implementations of "unreal" channel drivers, which lets
the unreal channel driver call specific functionality when it performs some
action. Currently, this is done to notify implementations when an
optimization operation has begun, and when an optimization operation has
succeeded.
* It adds Stasis-Core messages for Local channel bridging and Local channel
optimization. Local channel optimization is now two events: a Begin and an
End. Some consumers of Stasis-Core may want to know when an operation is
beginning so that they can 'prepare' their information; others will be more
concerned about when the operation has completed, so that they can 'fix up'
information. Stasis-Core allows for both, as does AMI.
Mark Michelson [Mon, 8 Jul 2013 13:57:28 +0000 (13:57 +0000)]
Fix some broken logic in sending outbound caller ID.
* trust_id_outbound was required even when the caller ID was not marked
private. This is against intentions and documentation.
* We now check both name and number privacy instead of checking name privacy
twice.
Matthew Jordan [Sun, 7 Jul 2013 21:29:40 +0000 (21:29 +0000)]
In a channel destructor dispose of items that raise Stasis message properly
This patch reorders certain actions that may raise Stasis messages in the
channel destructor such that they occur before the Stasis cache is cleared.
Once the Stasis cache is cleared, its rather a bad idea to be trying to
publish information about a channel.
(closes issue ASTERISK-22001)
Reported by: Jonathan Rose
Matthew Jordan [Sun, 7 Jul 2013 20:34:38 +0000 (20:34 +0000)]
Handle hangup logic in the Stasis message bus and consumers of Stasis messages
This patch does the following:
* It adds a new soft hangup flag AST_SOFTHANGUP_HANGUP_EXEC that is set when a
channel is executing dialplan hangup logic, i.e., the 'h' extension or a
hangup handler. Stasis messages now also convey the soft hangup flag so
consumers of the messages can know when a channel is executing said
hangup logic.
* It adds a new channel flag, AST_FLAG_DEAD, which is set when a channel is
well and truly dead. Not just a zombie, but dead, Jim. Manager, CEL, CDRs,
and other consumers of Stasis have been updated to look for this flag to
know when the channel should by lying six feet under.
* The CDR engine has been updated to better handle a channel entering and
leaving a bridge. Previously, a new CDR was automatically created when a
channel left a bridge and put into the 'Pending' state; however, this
way of handling CDRs made it difficult for the 'endbeforehexten' logic to
work correctly - there was always a new CDR waiting in the hangup logic
and, even if 'ended', wouldn't be the CDR people wanted to inspect in the
hangup routine. This patch completely removes the Pending state and instead
defers creation of the new CDR until it gets a new message that requires
a new CDR.
Matthew Jordan [Fri, 5 Jul 2013 17:33:33 +0000 (17:33 +0000)]
Refactor RTCP events over to Stasis; associate with channels
This patch does the following:
* It merges Jaco Kroon's patch from ASTERISK-20754, which provides channel
information in the RTCP events. Because Stasis provides a cache, Jaco's
patch was modified to pass the channel uniqueid to the RTP layer as
opposed to a pointer to the channel. This has the following benefits:
(1) It keeps the RTP engine 'clean' of references back to channels
(2) It prevents circular dependencies and other potential ref counting issues
* The RTP engine now allows any RTP implementation to raise RTCP messages.
Potentially, other implementations (such as res_rtp_multicast) could also
raise RTCP information. The engine provides structs to represent RTCP headers
and RTCP SR/RR reports.
* Some general refactoring in res_rtp_asterisk was done to try and tame the
RTCP code. It isn't perfect - that's *way* beyond the scope of this work -
but it does feel marginally better.
* A few random bugs were fixed in the RTCP statistics. (Example: performing an
assignment of a = a is probably not correct)
* We now raise RTCP events for each SR/RR sent/received. Previously we wouldn't
raise an event when we sent a RR report.
Note that this work will be of use to others who want to monitor call quality
or build modules that report call quality statistics. Since the events are now
moving across the Stasis message bus, this is far easier to accomplish. It is
also a first step (though by no means the last step) towards getting Olle's
pinefrog work incorporated.
Again: note that the patch by Jaco Kroon was modified slightly for this work;
however, he did all of the hard work in finding the right places to set the
channel in the RTP engine across the channel drivers. Much thanks goes to Jaco
for his hard work here.
Matthew Jordan [Fri, 5 Jul 2013 03:08:58 +0000 (03:08 +0000)]
Remove parkinglot from the channel snapshot
Legacy channel drivers often include the ability to set a default parking lot
on an endpoint basis; when channels are created for that endpoint, they inherit
the parkinglot option. Parking used to use this option more frequently; while
it is still supported, other options (such as using channel variables or
creation of a custom parkinglot) are supported. More importantly, conveying the
parkinglot information through a channel snapshot isn't terribly useful - it
is rarely (if ever) changed on a channel and some consumers of channel
snapshots, such as ARI, will never use the information.
(closes issue ASTERISK-21968)
Reported by: Matt Jordan
Jonathan Rose [Thu, 4 Jul 2013 18:46:56 +0000 (18:46 +0000)]
res_parking: Replace Parker snapshots with ParkerDialString
This process also involved a large amount of rework regarding how to redial
the Parker when a channel leaves a parking lot due to timeout. An attended
transfer channel variable has been added to attended transfers to extensions
that will eventually park (but haven't at the time of transfer) as well.
This resolves one of the two BUGBUG comments remaining in res_parking.
(issues ASTERISK-21877)
Reported by: Matt Jordan
Review: https://reviewboard.asterisk.org/r/2638/
Richard Mudgett [Wed, 3 Jul 2013 22:36:38 +0000 (22:36 +0000)]
OneTouchRecord: Make so Monitor/MixMonitor can be toggled/started/stopped.
The OneTouchRecord feature has historically been a toggle. This patch
adds the ability to make the OneTouchRecord hook optionally start/stop
recording only. If OneTouchRecord is already doing what is requested then
only the invoker hears the courtesy tone and/or start/stop recording
message.
The new feature is written so we could easily add explicit start/stop
recording DTMF hooks for Monitor and MixMonitor.
The majority of the changes in bridge_builtin_features.c is a refactoring
of the OneTouchRecord code (Monitor and MixMonitor versions) so it is easy
to direct the toggle/start/stop functionality.
Matthew Jordan [Wed, 3 Jul 2013 22:04:08 +0000 (22:04 +0000)]
Fix some bugs in CDRs; add some CLI commands to help debugging
This patch fixes a few minor bugs and one major one: the CDR by bridge
container was less than helpful. The mechanism previously used to try
and find all of the CDRs in a particular bridge ended up missing CDRs,
resulting in incorrect records.
When looking up CDRs in a bridge, we now just bite the bullet and do
a selection across all existing CDRs.
Mark Michelson [Wed, 3 Jul 2013 20:41:00 +0000 (20:41 +0000)]
Publish a bridge enter before pulling on a push-and-swap operation.
Prior to this patch, the order of procedures on a bridge push was
* Add new bridge channel to bridge's array.
* Pull the swap channel out of the bridge
* Publish a bridge enter event.
The problem is that when the swap channel was pulled from the bridge,
a bridge leave event would be published. The bridge snapshot
published during the bridge leave showed the new channel that had
been added to the bridge, but there had been no bridge enter event
for that channel.
The fix provided here was to change the order a bit
* Add new bridge channel to bridge's array.
* Publish bridge enter event.
* Pull the swap channel out of the bridge.
This makes it so that the bridge snapshots during the stasis
events are accurate.
David M. Lee [Wed, 3 Jul 2013 19:46:50 +0000 (19:46 +0000)]
Fix load errors related to the new ari_model_validators.
The Asterisk strategy of loading modules with RTLD_LAZY to extract metadata
from the module works well enough, until you try to take the address of a
function.
If a module takes the address of a function, that function needs to be
resolved at load time. That kinda defeats RTLD_LAZY.
This patch adds some ari_validator_{id}_fn() wrapper functions for safely
getting the function pointer from a different module.
David M. Lee [Wed, 3 Jul 2013 17:58:45 +0000 (17:58 +0000)]
ARI - channel recording support
This patch is the first step in adding recording support to the
Asterisk REST Interface.
Recordings are stored in /var/spool/recording. Since recordings may be
destructive (overwriting existing files), the API rejects attempts to
escape the recording directory (avoiding issues if someone attempts to
record to ../../lib/sounds/greeting, for example).
David M. Lee [Wed, 3 Jul 2013 16:33:13 +0000 (16:33 +0000)]
ARI authentication.
This patch adds authentication support to ARI.
Two authentication methods are supported. The first is HTTP Basic
authentication, as specified in RFC 2617[1]. The second is by simply
passing the username and password as an ?api_key query parameter
(which allows swagger-ui[2] to authenticate more easily).
ARI usernames and passwords are configured in the ari.conf file
(formerly known as stasis_http.conf). The user may be set to
`read_only`, which will prohibit the user from issuing POST, DELETE,
etc. Also, the user's password may be specified in either plaintext,
or encrypted using the crypt() function.
Several other notes about the patch.
* A few command line commands for seeing ARI config and status were
also added.
* The configuration parsing grew big enough that I extracted it to
its own file.
The event discriminator was the field name instead of being a value in
the object, due to limitations in how Swagger 1.1 could model objects.
While technically sufficient in communicating event information, it was
really difficult to deal with in terms of client side JSON handling.
This patch takes advantage of a proposed extension[1] to Swagger which
allows type variance through the use of a discriminator field. This had
a domino effect that made this a surprisingly large patch.
In changing the models, I also had to change the swagger_model.py
processor so it can handle the type discriminator and subtyping. I took
that a big step forward, and using that information to generate an
ari_model module, which can validate a JSON object against the Swagger
model.
The REST and WebSocket generators were changed to take advantage of the
validators. If compiled with AST_DEVMODE enabled, JSON objects that
don't match their corresponding models will not be sent out. For REST
API calls, a 500 Internal Server response is sent. For WebSockets, the
invalid JSON message is replaced with an error message.
Since this took over about half of the job of the existing JSON
generators, and the .to_json virtual function on messages took over the
other half, I reluctantly removed the generators.
The validators turned up all sorts of errors and inconsistencies in our
data models, and the code. These were cleaned up, with checks in the
code generator avoid some of the consistency problems in the future.
* The model for a channel snapshot was trimmed down to match the
information sent via AMI. Many of the field being sent were not
useful in the general case.
* The model for a bridge snapshot was updated to be more consistent
with the other ARI models.
Another impact of introducing subtyping was that the swagger-codegen
documentation generator was insufficient (at least until it catches up
with Swagger 1.2). I wanted it to be easier to generate docs for the API
anyways, so I ported the wiki pages to use the Asterisk Swagger
generator. In the process, I was able to clean up many of the model
links, which would occasionally give inconsistent results on the wiki. I
also added error responses to the wiki docs, making the wiki
documentation more complete.
Finally, since Stasis-HTTP will now be named Asterisk REST Interface
(ARI), any new functions and files I created carry the ari_ prefix. I
changed a few stasis_http references to ari where it was non-intrusive
and made sense.
David M. Lee [Wed, 3 Jul 2013 16:32:00 +0000 (16:32 +0000)]
Shuffle RESTful URL's around.
This patch moves the RESTful URL's around to more appropriate
locations for release.
The /stasis URL's are moved to /ari, since Asterisk REST Interface was
a more appropriate name than Stasis-HTTP. (Most of the code still has
stasis_http references, but they will be cleaned up after there are no
more outstanding branches that would have merge conflicts with such a
change).
A larger change was moving the ARI events WebSocket off of the shared
/ws URL to its permanent home on /ari/events. The Swagger code
generator was extended to handle "upgrade: websocket" and
"websocketProtocol:" attributes on an operation.
The WebSocket module was modified to better handle WebSocket servers
that have a single registered protocol handler. If a client
connections does not specify the Sec-WebSocket-Protocol header, and
the server has a single protocol handler registered, the WebSocket
server will go ahead and accept the client for that subprotocol.
Kevin Harwell [Tue, 2 Jul 2013 17:20:20 +0000 (17:20 +0000)]
Stasis - Refactor AOC Events
Refactored the AMI events in AOC onto Stasis-Core. The ast_aoc_manager_event
function now publishes a channel snapshot, along with a JSON blob describing
the advice of charge. A "to_ami" handler has also been added that converts
the channel snapshot and AOC event data back into the appropriate data structure
for use with AMI.
(closes issue ASTERISK-21472)
Reported by: Matt Jordan
Review: https://reviewboard.asterisk.org/r/2643/
Kevin Harwell [Tue, 2 Jul 2013 17:06:06 +0000 (17:06 +0000)]
New SIP Channel driver: Always Auth Reject
If no matching endpoint is found for the incoming request Asterisk will respond
with a 401 Unauthorized (rejecting the request), but will first challenge if
no authorization creditials are given.
Changes also included moving ACL options into a new global 'security'
configuration section in res_sip.conf.
(closes issue ASTERISK-21433)
Reported by: Matt Jordan
Review: https://reviewboard.asterisk.org/r/2554/
Kevin Harwell [Mon, 1 Jul 2013 21:28:32 +0000 (21:28 +0000)]
New SIP Channel Driver - Add CLI/AMI initiated NOTIFY requests
Added the ability to send unsolicited NOTIFY requests to a particular endpoint
with a configured payload. Added both CLI and AMI support. For a given
endpoint, this module will iterate over all its contacts sending the appropriate
NOTIFY request to each.
(closes issue ASTERISK-21436)
Reported by: Matt Jordan
Review: https://reviewboard.asterisk.org/r/2623/
Matthew Jordan [Mon, 1 Jul 2013 21:24:20 +0000 (21:24 +0000)]
Prevent crash during synchronous AMI origination by ref bumping returned channel
The originate APIs allow callers to provide a pointer to a channel that will
point to the originated channel if the function call succeeds. This is used by AMI
to provide channel information when the originate is performed synchronously.
Unfortunately, if the originate fails in certain ways, the outbound channel is
already disposed of during the dialing itself. This results in the channel being
improperly dereferenced by the internal originate function in pbx.c.
This patch ref bumps the channel to prevent this from occurring. Callers must now
unlock and unref the channel (which is more in line with general channel management
guidelines anyway).
This only affects manager, as it is the only consumer of this API function that
actually passes in a channel pointer.
Jonathan Rose [Mon, 1 Jul 2013 16:01:24 +0000 (16:01 +0000)]
bridge_features: Support One touch Monitor/MixMonitor
In addition to porting those features, they now enjoy greater feature parity
with one another. Specifically, AutoMixMon now has a start and stop
message that can be specified with TOUCH_MIXMONITOR_MESSAGE_START and
TOUCH_MIXMONITOR_MESSAGE_STOP.
(closes issue ASTERISK-21553)
Reported by: Matt Jordan
Review: https://reviewboard.asterisk.org/r/2620/
Matthew Jordan [Fri, 28 Jun 2013 17:31:33 +0000 (17:31 +0000)]
Handle an originated channel being sent into a non-empty bridge
Originated channels are a bit odd - they are technically a dialed channel (thus
the party B or peer) but, since there is no caller, they are treated as the
party A. When entering into a bridge that already contains participants, the CDR
engine - if the CDR record is in the Dial state - attempts to match the person
entering the bridge with an existing participant. The idea is that if you dialed
someone and the person you dialed is already in the bridge, you don't need a new
CDR record, the existing CDR record describes the relationship.
Unfortunately, for an originated channel, there is no Party B. If no one was in
the bridge this didn't cause any issues; however, if participants were in the
bridge the CDR engine would attempt to match a non-existant Party B on the
channel's CDR record and explode.
This patch fixes that, and a unit test has been added to cover this case.
Matthew Jordan [Fri, 28 Jun 2013 15:50:56 +0000 (15:50 +0000)]
Better handle parking in CDRs
Parking typically occurs when a channel is transferred to a parking extension.
When this occurs, the channel never actually hits the dialplan if the extension
it was transferred to was a "parking extension", that is, the extension in
the first priority calls the Park application. Instead, the channel is
immediately sent into the holding bridge acting as the parking bridge.
This is problematic.
Because we never go out to the dialplan, the CDRs won't transition properly
and the application field will not be set to "Park". CDRs typically swallow
holding bridges, so the CDR itself won't even be generated.
This patch handles this by pulling out the holding bridge handling into its
own CDR state. CDRs now have an explicit parking state that accounts for this
specific subclass of the holding bridge. In addition, we handle the parking
stasis message to set application specific data on the CDR such that the
last known application for the CDR properly reflects "Park".
This is a bit sad since we're working around the odd internal implementation
of parking that exists in Asterisk (and that we had to maintain in order to
continue to meet some odd use cases of parking), but at least the code to
handle that is where it belongs: in CDRs as opposed to sprinkled liberally
throughout the codebase.
This patch also properly clears the OUTBOUND channel flag from a channel when
it leaves a bridge, and tweaks up dialing handling to properly compare the
correct CDR with the channel calling/being dialed.
David M. Lee [Fri, 28 Jun 2013 02:14:27 +0000 (02:14 +0000)]
Removed stray apostrophe.
Apparently the pluralization of an acronym does not use an apostophe,
according to most modern style guides. I feel like I've been living a
lie this whole time.
David M. Lee [Fri, 28 Jun 2013 01:07:32 +0000 (01:07 +0000)]
Removed the automatic 302 redirects for ARI URL's that end with a slash.
There were some problems redirecting RESTful API requests; notably the client
would change the request method to GET on the redirected requests. After some
looking into, I decided that a 404 would be simpler and have more consistent
behavior.
Add config framework OPT_CHAR_ARRAY_T and OPT_STRINGFIELD_T non-empty
requirement option. There are cases were you don't want a config option
string to be empty. To require the option string to be non-empty, just
set the aco_option_register() flags parameter to non-zero.
* Updated some config framework enum aco_option_type comments.
Richard Mudgett [Wed, 26 Jun 2013 14:38:57 +0000 (14:38 +0000)]
Fix several problems with ast_bridge_add_channel().
* Fix locking problems. ast_bridge_move() locks two bridges. To do that,
deadlock avoidance must be done. Called bridge_move_locked() instead.
* Fix inconsistency in the bridge dissolve check callers. The original
caller has already removed the channel from the bridge. The new caller
has not removed the channel from the bridge. Reverted
bridge_dissolve_check() and added bridge_dissolve_check_stolen() to be
used by the new caller on the original bridge after the channel is moved
to the new bridge.
* Fix memory leak of features if the added channel was already in a
bridge.
Richard Mudgett [Wed, 26 Jun 2013 01:46:30 +0000 (01:46 +0000)]
Fix incorrect calls to ast_bridge_impart().
There was a misunderstanding about ast_bridge_impart()'s handling of the
imparted channel's reference. The channel reference is passed by the
caller unless ast_bridge_impart() returns an error.
* Fixed a memory leak in conf_announce_channel_push() if the impart
failed.
Matthew Jordan [Tue, 25 Jun 2013 01:12:58 +0000 (01:12 +0000)]
Fix memory/ref counting leaks in a variety of locations
This patch fixes the following memory leaks:
* http.c: The structure containing the addresses to bind to was not being
deallocated when no longer used
* named_acl.c: The global configuration information was not disposed of
* config_options.c: An invalid read was occurring for certain option types.
* res_calendar.c: The loaded calendars on module unload were not being
properly disposed of.
* chan_motif.c: The format capabilities needed to be disposed of on module
unload. In addition, this now specifies the default options for the
maxpayloads and maxicecandidates in such a way that it doesn't cause the
invalid read in config_options.c to occur.
(issue ASTERISK-21906)
Reported by: John Hardin
patches:
http.patch uploaded by jhardin (license 6512)
named_acl.patch uploaded by jhardin (license 6512)
config_options.patch uploaded by jhardin (license 6512)
res_calendar.patch uploaded by jhardin (license 6512)
chan_motif.patch uploaded by jhardin (license 6512)
........
Merged revisions 392810 from http://svn.asterisk.org/svn/asterisk/branches/11
Matthew Jordan [Mon, 24 Jun 2013 23:56:54 +0000 (23:56 +0000)]
Fix a variety of memory leaks
This patch addresses the following memory/ref counting leaks:
* main/devicestate.c - unsubscribe and join our devicestate message
subscription
* main/cel.c - clean up the datastore and config objects on exist
* main/parking.c - cleanup memory leak of retriever snapshot on message
payload destruction
* res/parking/parking_bridge.c - cleanup memory leak of retrieve snapshot
on message payload destruction
* main/presencestate.c - unsubscribe and join the caching topic on exit
* manager.c - properly unregister the manager action "BlindTransfer"
* sorcery.c - shutdown the threadpool on exit and dispose of any wizards
(issue ASTERISK-21906)
Reported by: John Hardin
patches:
cel.patch uploaded by jhardin (license #6512)
devicestate.patch uploaded by jhardin (license #6512)
manager.patch uploaded by jardin (license #6512)
presencestate.patch uploaded by jhardin (license #6512)
retriever-channel-snapshot.patch uploaded by jhardin (license #6512)
sorcery.patch uploaded by jhardin (license #6512)
projects / thirdparty / asterisk.git / log
