asterisk/main/manager_channels.c

1455 lines
44 KiB
C
Raw Normal View History

Move more channel events to Stasis; move res_json.c to main/json.c. This patch started out simply as fixing the bouncing tests introduced in r382685, but required some other changes to give it a decent implementation. To fix the bouncing tests, the UserEvent and Newexten AMI events needed to be refactored to dispatch via Stasis. Dispatching directly to AMI resulted in those events sometimes getting ahead of the associated Newchannel events, which would understandably confuse anyone. I found that instead of creating a zillion different message types and structures associated with them, it would be preferable to define a message type that has a channel snapshot and a blob of structured data with a small bit of additional information. The JSON object model provides a very nice way of representing structured data, so I went with that. * Move JSON support from res_json.c to main/json.c * Made libjansson-dev a required dependency * Added an ast_channel_blob message type, which has a channel snapshot and JSON blob of data. * Changed UserEvent and Newexten events so that they are dispatched via ast_channel_blob messages on the channel's topic. * Got rid of the ast_channel_varset message; used ast_channel_blob instead. * Extracted the manager functions converting Stasis channel events to AMI events into manager_channel.c. (issue ASTERISK-21096) Review: https://reviewboard.asterisk.org/r/2381/ git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@383579 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2013-03-22 14:06:46 +00:00
/*
* Asterisk -- An open source telephony toolkit.
*
* Copyright (C) 2013, Digium, Inc.
*
* David M. Lee, II <dlee@digium.com>
*
* See http://www.asterisk.org for more information about
* the Asterisk project. Please do not directly contact
* any of the maintainers of this project for assistance;
* the project provides a web site, mailing lists and IRC
* channels for your use.
*
* This program is free software, distributed under the terms of
* the GNU General Public License Version 2. See the LICENSE file
* at the top of the source tree.
*/
/*! \file
*
* \brief The Asterisk Management Interface - AMI (channel event handling)
*
* \author David M. Lee, II <dlee@digium.com>
*
* AMI generated many per-channel and global-channel events by converting Stasis
* messages to AMI events. It makes sense to simply put them into a single file.
*/
#include "asterisk.h"
#include "asterisk/callerid.h"
Move more channel events to Stasis; move res_json.c to main/json.c. This patch started out simply as fixing the bouncing tests introduced in r382685, but required some other changes to give it a decent implementation. To fix the bouncing tests, the UserEvent and Newexten AMI events needed to be refactored to dispatch via Stasis. Dispatching directly to AMI resulted in those events sometimes getting ahead of the associated Newchannel events, which would understandably confuse anyone. I found that instead of creating a zillion different message types and structures associated with them, it would be preferable to define a message type that has a channel snapshot and a blob of structured data with a small bit of additional information. The JSON object model provides a very nice way of representing structured data, so I went with that. * Move JSON support from res_json.c to main/json.c * Made libjansson-dev a required dependency * Added an ast_channel_blob message type, which has a channel snapshot and JSON blob of data. * Changed UserEvent and Newexten events so that they are dispatched via ast_channel_blob messages on the channel's topic. * Got rid of the ast_channel_varset message; used ast_channel_blob instead. * Extracted the manager functions converting Stasis channel events to AMI events into manager_channel.c. (issue ASTERISK-21096) Review: https://reviewboard.asterisk.org/r/2381/ git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@383579 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2013-03-22 14:06:46 +00:00
#include "asterisk/channel.h"
#include "asterisk/manager.h"
#include "asterisk/stasis_message_router.h"
#include "asterisk/pbx.h"
#include "asterisk/stasis_channels.h"
Move more channel events to Stasis; move res_json.c to main/json.c. This patch started out simply as fixing the bouncing tests introduced in r382685, but required some other changes to give it a decent implementation. To fix the bouncing tests, the UserEvent and Newexten AMI events needed to be refactored to dispatch via Stasis. Dispatching directly to AMI resulted in those events sometimes getting ahead of the associated Newchannel events, which would understandably confuse anyone. I found that instead of creating a zillion different message types and structures associated with them, it would be preferable to define a message type that has a channel snapshot and a blob of structured data with a small bit of additional information. The JSON object model provides a very nice way of representing structured data, so I went with that. * Move JSON support from res_json.c to main/json.c * Made libjansson-dev a required dependency * Added an ast_channel_blob message type, which has a channel snapshot and JSON blob of data. * Changed UserEvent and Newexten events so that they are dispatched via ast_channel_blob messages on the channel's topic. * Got rid of the ast_channel_varset message; used ast_channel_blob instead. * Extracted the manager functions converting Stasis channel events to AMI events into manager_channel.c. (issue ASTERISK-21096) Review: https://reviewboard.asterisk.org/r/2381/ git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@383579 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2013-03-22 14:06:46 +00:00
/*** DOCUMENTATION
<managerEvent language="en_US" name="Newchannel">
<managerEventInstance class="EVENT_FLAG_CALL">
<synopsis>Raised when a new channel is created.</synopsis>
<syntax>
<channel_snapshot/>
</syntax>
<see-also>
<ref type="managerEvent">Newstate</ref>
<ref type="managerEvent">Hangup</ref>
</see-also>
</managerEventInstance>
</managerEvent>
<managerEvent language="en_US" name="Newstate">
<managerEventInstance class="EVENT_FLAG_CALL">
<synopsis>Raised when a channel's state changes.</synopsis>
<syntax>
<channel_snapshot/>
</syntax>
<see-also>
<ref type="managerEvent">Newchannel</ref>
<ref type="managerEvent">Hangup</ref>
</see-also>
</managerEventInstance>
</managerEvent>
<managerEvent language="en_US" name="Hangup">
<managerEventInstance class="EVENT_FLAG_CALL">
<synopsis>Raised when a channel is hung up.</synopsis>
<syntax>
<channel_snapshot/>
Move more channel events to Stasis; move res_json.c to main/json.c. This patch started out simply as fixing the bouncing tests introduced in r382685, but required some other changes to give it a decent implementation. To fix the bouncing tests, the UserEvent and Newexten AMI events needed to be refactored to dispatch via Stasis. Dispatching directly to AMI resulted in those events sometimes getting ahead of the associated Newchannel events, which would understandably confuse anyone. I found that instead of creating a zillion different message types and structures associated with them, it would be preferable to define a message type that has a channel snapshot and a blob of structured data with a small bit of additional information. The JSON object model provides a very nice way of representing structured data, so I went with that. * Move JSON support from res_json.c to main/json.c * Made libjansson-dev a required dependency * Added an ast_channel_blob message type, which has a channel snapshot and JSON blob of data. * Changed UserEvent and Newexten events so that they are dispatched via ast_channel_blob messages on the channel's topic. * Got rid of the ast_channel_varset message; used ast_channel_blob instead. * Extracted the manager functions converting Stasis channel events to AMI events into manager_channel.c. (issue ASTERISK-21096) Review: https://reviewboard.asterisk.org/r/2381/ git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@383579 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2013-03-22 14:06:46 +00:00
<parameter name="Cause">
<para>A numeric cause code for why the channel was hung up.</para>
</parameter>
<parameter name="Cause-txt">
<para>A description of why the channel was hung up.</para>
</parameter>
</syntax>
<see-also>
<ref type="managerEvent">Newchannel</ref>
<ref type="managerEvent">SoftHangupRequest</ref>
<ref type="managerEvent">HangupRequest</ref>
<ref type="managerEvent">Newstate</ref>
</see-also>
Move more channel events to Stasis; move res_json.c to main/json.c. This patch started out simply as fixing the bouncing tests introduced in r382685, but required some other changes to give it a decent implementation. To fix the bouncing tests, the UserEvent and Newexten AMI events needed to be refactored to dispatch via Stasis. Dispatching directly to AMI resulted in those events sometimes getting ahead of the associated Newchannel events, which would understandably confuse anyone. I found that instead of creating a zillion different message types and structures associated with them, it would be preferable to define a message type that has a channel snapshot and a blob of structured data with a small bit of additional information. The JSON object model provides a very nice way of representing structured data, so I went with that. * Move JSON support from res_json.c to main/json.c * Made libjansson-dev a required dependency * Added an ast_channel_blob message type, which has a channel snapshot and JSON blob of data. * Changed UserEvent and Newexten events so that they are dispatched via ast_channel_blob messages on the channel's topic. * Got rid of the ast_channel_varset message; used ast_channel_blob instead. * Extracted the manager functions converting Stasis channel events to AMI events into manager_channel.c. (issue ASTERISK-21096) Review: https://reviewboard.asterisk.org/r/2381/ git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@383579 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2013-03-22 14:06:46 +00:00
</managerEventInstance>
</managerEvent>
<managerEvent language="en_US" name="HangupRequest">
Move more channel events to Stasis; move res_json.c to main/json.c. This patch started out simply as fixing the bouncing tests introduced in r382685, but required some other changes to give it a decent implementation. To fix the bouncing tests, the UserEvent and Newexten AMI events needed to be refactored to dispatch via Stasis. Dispatching directly to AMI resulted in those events sometimes getting ahead of the associated Newchannel events, which would understandably confuse anyone. I found that instead of creating a zillion different message types and structures associated with them, it would be preferable to define a message type that has a channel snapshot and a blob of structured data with a small bit of additional information. The JSON object model provides a very nice way of representing structured data, so I went with that. * Move JSON support from res_json.c to main/json.c * Made libjansson-dev a required dependency * Added an ast_channel_blob message type, which has a channel snapshot and JSON blob of data. * Changed UserEvent and Newexten events so that they are dispatched via ast_channel_blob messages on the channel's topic. * Got rid of the ast_channel_varset message; used ast_channel_blob instead. * Extracted the manager functions converting Stasis channel events to AMI events into manager_channel.c. (issue ASTERISK-21096) Review: https://reviewboard.asterisk.org/r/2381/ git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@383579 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2013-03-22 14:06:46 +00:00
<managerEventInstance class="EVENT_FLAG_CALL">
<synopsis>Raised when a hangup is requested.</synopsis>
Move more channel events to Stasis; move res_json.c to main/json.c. This patch started out simply as fixing the bouncing tests introduced in r382685, but required some other changes to give it a decent implementation. To fix the bouncing tests, the UserEvent and Newexten AMI events needed to be refactored to dispatch via Stasis. Dispatching directly to AMI resulted in those events sometimes getting ahead of the associated Newchannel events, which would understandably confuse anyone. I found that instead of creating a zillion different message types and structures associated with them, it would be preferable to define a message type that has a channel snapshot and a blob of structured data with a small bit of additional information. The JSON object model provides a very nice way of representing structured data, so I went with that. * Move JSON support from res_json.c to main/json.c * Made libjansson-dev a required dependency * Added an ast_channel_blob message type, which has a channel snapshot and JSON blob of data. * Changed UserEvent and Newexten events so that they are dispatched via ast_channel_blob messages on the channel's topic. * Got rid of the ast_channel_varset message; used ast_channel_blob instead. * Extracted the manager functions converting Stasis channel events to AMI events into manager_channel.c. (issue ASTERISK-21096) Review: https://reviewboard.asterisk.org/r/2381/ git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@383579 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2013-03-22 14:06:46 +00:00
<syntax>
<channel_snapshot/>
<xi:include xpointer="xpointer(/docs/managerEvent[@name='Hangup']/managerEventInstance/syntax/parameter[@name='Cause'])" />
Move more channel events to Stasis; move res_json.c to main/json.c. This patch started out simply as fixing the bouncing tests introduced in r382685, but required some other changes to give it a decent implementation. To fix the bouncing tests, the UserEvent and Newexten AMI events needed to be refactored to dispatch via Stasis. Dispatching directly to AMI resulted in those events sometimes getting ahead of the associated Newchannel events, which would understandably confuse anyone. I found that instead of creating a zillion different message types and structures associated with them, it would be preferable to define a message type that has a channel snapshot and a blob of structured data with a small bit of additional information. The JSON object model provides a very nice way of representing structured data, so I went with that. * Move JSON support from res_json.c to main/json.c * Made libjansson-dev a required dependency * Added an ast_channel_blob message type, which has a channel snapshot and JSON blob of data. * Changed UserEvent and Newexten events so that they are dispatched via ast_channel_blob messages on the channel's topic. * Got rid of the ast_channel_varset message; used ast_channel_blob instead. * Extracted the manager functions converting Stasis channel events to AMI events into manager_channel.c. (issue ASTERISK-21096) Review: https://reviewboard.asterisk.org/r/2381/ git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@383579 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2013-03-22 14:06:46 +00:00
</syntax>
<see-also>
<ref type="managerEvent">SoftHangupRequest</ref>
<ref type="managerEvent">Hangup</ref>
</see-also>
Move more channel events to Stasis; move res_json.c to main/json.c. This patch started out simply as fixing the bouncing tests introduced in r382685, but required some other changes to give it a decent implementation. To fix the bouncing tests, the UserEvent and Newexten AMI events needed to be refactored to dispatch via Stasis. Dispatching directly to AMI resulted in those events sometimes getting ahead of the associated Newchannel events, which would understandably confuse anyone. I found that instead of creating a zillion different message types and structures associated with them, it would be preferable to define a message type that has a channel snapshot and a blob of structured data with a small bit of additional information. The JSON object model provides a very nice way of representing structured data, so I went with that. * Move JSON support from res_json.c to main/json.c * Made libjansson-dev a required dependency * Added an ast_channel_blob message type, which has a channel snapshot and JSON blob of data. * Changed UserEvent and Newexten events so that they are dispatched via ast_channel_blob messages on the channel's topic. * Got rid of the ast_channel_varset message; used ast_channel_blob instead. * Extracted the manager functions converting Stasis channel events to AMI events into manager_channel.c. (issue ASTERISK-21096) Review: https://reviewboard.asterisk.org/r/2381/ git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@383579 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2013-03-22 14:06:46 +00:00
</managerEventInstance>
</managerEvent>
<managerEvent language="en_US" name="SoftHangupRequest">
Move more channel events to Stasis; move res_json.c to main/json.c. This patch started out simply as fixing the bouncing tests introduced in r382685, but required some other changes to give it a decent implementation. To fix the bouncing tests, the UserEvent and Newexten AMI events needed to be refactored to dispatch via Stasis. Dispatching directly to AMI resulted in those events sometimes getting ahead of the associated Newchannel events, which would understandably confuse anyone. I found that instead of creating a zillion different message types and structures associated with them, it would be preferable to define a message type that has a channel snapshot and a blob of structured data with a small bit of additional information. The JSON object model provides a very nice way of representing structured data, so I went with that. * Move JSON support from res_json.c to main/json.c * Made libjansson-dev a required dependency * Added an ast_channel_blob message type, which has a channel snapshot and JSON blob of data. * Changed UserEvent and Newexten events so that they are dispatched via ast_channel_blob messages on the channel's topic. * Got rid of the ast_channel_varset message; used ast_channel_blob instead. * Extracted the manager functions converting Stasis channel events to AMI events into manager_channel.c. (issue ASTERISK-21096) Review: https://reviewboard.asterisk.org/r/2381/ git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@383579 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2013-03-22 14:06:46 +00:00
<managerEventInstance class="EVENT_FLAG_CALL">
<synopsis>Raised when a soft hangup is requested with a specific cause code.</synopsis>
Move more channel events to Stasis; move res_json.c to main/json.c. This patch started out simply as fixing the bouncing tests introduced in r382685, but required some other changes to give it a decent implementation. To fix the bouncing tests, the UserEvent and Newexten AMI events needed to be refactored to dispatch via Stasis. Dispatching directly to AMI resulted in those events sometimes getting ahead of the associated Newchannel events, which would understandably confuse anyone. I found that instead of creating a zillion different message types and structures associated with them, it would be preferable to define a message type that has a channel snapshot and a blob of structured data with a small bit of additional information. The JSON object model provides a very nice way of representing structured data, so I went with that. * Move JSON support from res_json.c to main/json.c * Made libjansson-dev a required dependency * Added an ast_channel_blob message type, which has a channel snapshot and JSON blob of data. * Changed UserEvent and Newexten events so that they are dispatched via ast_channel_blob messages on the channel's topic. * Got rid of the ast_channel_varset message; used ast_channel_blob instead. * Extracted the manager functions converting Stasis channel events to AMI events into manager_channel.c. (issue ASTERISK-21096) Review: https://reviewboard.asterisk.org/r/2381/ git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@383579 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2013-03-22 14:06:46 +00:00
<syntax>
<channel_snapshot/>
<xi:include xpointer="xpointer(/docs/managerEvent[@name='Hangup']/managerEventInstance/syntax/parameter[@name='Cause'])" />
</syntax>
<see-also>
<ref type="managerEvent">HangupRequest</ref>
<ref type="managerEvent">Hangup</ref>
</see-also>
</managerEventInstance>
</managerEvent>
<managerEvent language="en_US" name="NewExten">
<managerEventInstance class="EVENT_FLAG_DIALPLAN">
<synopsis>Raised when a channel enters a new context, extension, priority.</synopsis>
<syntax>
<channel_snapshot/>
<parameter name="Extension">
<para>Deprecated in 12, but kept for
backward compatability. Please use
'Exten' instead.</para>
</parameter>
<parameter name="Application">
<para>The application about to be executed.</para>
</parameter>
<parameter name="AppData">
<para>The data to be passed to the application.</para>
</parameter>
</syntax>
</managerEventInstance>
</managerEvent>
<managerEvent language="en_US" name="NewCallerid">
<managerEventInstance class="EVENT_FLAG_CALL">
<synopsis>Raised when a channel receives new Caller ID information.</synopsis>
<syntax>
<channel_snapshot/>
<parameter name="CID-CallingPres">
<para>A description of the Caller ID presentation.</para>
</parameter>
Move more channel events to Stasis; move res_json.c to main/json.c. This patch started out simply as fixing the bouncing tests introduced in r382685, but required some other changes to give it a decent implementation. To fix the bouncing tests, the UserEvent and Newexten AMI events needed to be refactored to dispatch via Stasis. Dispatching directly to AMI resulted in those events sometimes getting ahead of the associated Newchannel events, which would understandably confuse anyone. I found that instead of creating a zillion different message types and structures associated with them, it would be preferable to define a message type that has a channel snapshot and a blob of structured data with a small bit of additional information. The JSON object model provides a very nice way of representing structured data, so I went with that. * Move JSON support from res_json.c to main/json.c * Made libjansson-dev a required dependency * Added an ast_channel_blob message type, which has a channel snapshot and JSON blob of data. * Changed UserEvent and Newexten events so that they are dispatched via ast_channel_blob messages on the channel's topic. * Got rid of the ast_channel_varset message; used ast_channel_blob instead. * Extracted the manager functions converting Stasis channel events to AMI events into manager_channel.c. (issue ASTERISK-21096) Review: https://reviewboard.asterisk.org/r/2381/ git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@383579 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2013-03-22 14:06:46 +00:00
</syntax>
<see-also>
<ref type="function">CALLERID</ref>
</see-also>
Move more channel events to Stasis; move res_json.c to main/json.c. This patch started out simply as fixing the bouncing tests introduced in r382685, but required some other changes to give it a decent implementation. To fix the bouncing tests, the UserEvent and Newexten AMI events needed to be refactored to dispatch via Stasis. Dispatching directly to AMI resulted in those events sometimes getting ahead of the associated Newchannel events, which would understandably confuse anyone. I found that instead of creating a zillion different message types and structures associated with them, it would be preferable to define a message type that has a channel snapshot and a blob of structured data with a small bit of additional information. The JSON object model provides a very nice way of representing structured data, so I went with that. * Move JSON support from res_json.c to main/json.c * Made libjansson-dev a required dependency * Added an ast_channel_blob message type, which has a channel snapshot and JSON blob of data. * Changed UserEvent and Newexten events so that they are dispatched via ast_channel_blob messages on the channel's topic. * Got rid of the ast_channel_varset message; used ast_channel_blob instead. * Extracted the manager functions converting Stasis channel events to AMI events into manager_channel.c. (issue ASTERISK-21096) Review: https://reviewboard.asterisk.org/r/2381/ git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@383579 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2013-03-22 14:06:46 +00:00
</managerEventInstance>
</managerEvent>
<managerEvent language="en_US" name="NewConnectedLine">
<managerEventInstance class="EVENT_FLAG_CALL">
<synopsis>Raised when a channel's connected line information is changed.</synopsis>
<syntax>
<channel_snapshot/>
</syntax>
<see-also>
<ref type="function">CONNECTEDLINE</ref>
</see-also>
</managerEventInstance>
</managerEvent>
Update Asterisk's CDRs for the new bridging framework This patch is the initial push to update Asterisk's CDR engine for the new bridging framework. This patch guts the existing CDR engine and builds the new on top of messages coming across Stasis. As changes in channel state and bridge state are detected, CDRs are built and dispatched accordingly. This fundamentally changes CDRs in a few ways. (1) CDRs are now *very* reflective of the actual state of channels and bridges. This means CDRs track well with what an actual channel is doing - which is useful in transfer scenarios (which were previously difficult to pin down). It does, however, mean that CDRs cannot be 'fooled'. Previous behavior in Asterisk allowed for CDR applications, channels, and other properties to be spoofed in parts of the code - this no longer works. (2) CDRs have defined behavior in multi-party scenarios. This behavior will not be what everyone wants, but it is a defined behavior and as such, it is predictable. (3) The CDR manipulation functions and applications have been overhauled. Major changes have been made to ResetCDR and ForkCDR in particular. Many of the options for these two applications no longer made any sense with the new framework and the (slightly) more immutable nature of CDRs. There are a plethora of other changes. For a full description of CDR behavior, see the CDR specification on the Asterisk wiki. (closes issue ASTERISK-21196) Review: https://reviewboard.asterisk.org/r/2486/ git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@391947 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2013-06-17 03:00:38 +00:00
<managerEvent language="en_US" name="NewAccountCode">
<managerEventInstance class="EVENT_FLAG_CALL">
<synopsis>Raised when a Channel's AccountCode is changed.</synopsis>
<syntax>
<channel_snapshot/>
Update Asterisk's CDRs for the new bridging framework This patch is the initial push to update Asterisk's CDR engine for the new bridging framework. This patch guts the existing CDR engine and builds the new on top of messages coming across Stasis. As changes in channel state and bridge state are detected, CDRs are built and dispatched accordingly. This fundamentally changes CDRs in a few ways. (1) CDRs are now *very* reflective of the actual state of channels and bridges. This means CDRs track well with what an actual channel is doing - which is useful in transfer scenarios (which were previously difficult to pin down). It does, however, mean that CDRs cannot be 'fooled'. Previous behavior in Asterisk allowed for CDR applications, channels, and other properties to be spoofed in parts of the code - this no longer works. (2) CDRs have defined behavior in multi-party scenarios. This behavior will not be what everyone wants, but it is a defined behavior and as such, it is predictable. (3) The CDR manipulation functions and applications have been overhauled. Major changes have been made to ResetCDR and ForkCDR in particular. Many of the options for these two applications no longer made any sense with the new framework and the (slightly) more immutable nature of CDRs. There are a plethora of other changes. For a full description of CDR behavior, see the CDR specification on the Asterisk wiki. (closes issue ASTERISK-21196) Review: https://reviewboard.asterisk.org/r/2486/ git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@391947 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2013-06-17 03:00:38 +00:00
<parameter name="OldAccountCode">
<para>The channel's previous account code</para>
</parameter>
</syntax>
<see-also>
<ref type="function">CHANNEL</ref>
</see-also>
Update Asterisk's CDRs for the new bridging framework This patch is the initial push to update Asterisk's CDR engine for the new bridging framework. This patch guts the existing CDR engine and builds the new on top of messages coming across Stasis. As changes in channel state and bridge state are detected, CDRs are built and dispatched accordingly. This fundamentally changes CDRs in a few ways. (1) CDRs are now *very* reflective of the actual state of channels and bridges. This means CDRs track well with what an actual channel is doing - which is useful in transfer scenarios (which were previously difficult to pin down). It does, however, mean that CDRs cannot be 'fooled'. Previous behavior in Asterisk allowed for CDR applications, channels, and other properties to be spoofed in parts of the code - this no longer works. (2) CDRs have defined behavior in multi-party scenarios. This behavior will not be what everyone wants, but it is a defined behavior and as such, it is predictable. (3) The CDR manipulation functions and applications have been overhauled. Major changes have been made to ResetCDR and ForkCDR in particular. Many of the options for these two applications no longer made any sense with the new framework and the (slightly) more immutable nature of CDRs. There are a plethora of other changes. For a full description of CDR behavior, see the CDR specification on the Asterisk wiki. (closes issue ASTERISK-21196) Review: https://reviewboard.asterisk.org/r/2486/ git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@391947 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2013-06-17 03:00:38 +00:00
</managerEventInstance>
</managerEvent>
<managerEvent language="en_US" name="DialBegin">
<managerEventInstance class="EVENT_FLAG_CALL">
<synopsis>Raised when a dial action has started.</synopsis>
<syntax>
<channel_snapshot/>
<channel_snapshot prefix="Dest"/>
<parameter name="DialString">
<para>The non-technology specific device being dialed.</para>
</parameter>
</syntax>
<see-also>
<ref type="application">Dial</ref>
<ref type="application">Originate</ref>
<ref type="manager">Originate</ref>
<ref type="managerEvent">DialEnd</ref>
</see-also>
</managerEventInstance>
</managerEvent>
<managerEvent language="en_US" name="DialState">
<managerEventInstance class="EVENT_FLAG_CALL">
<synopsis>Raised when dial status has changed.</synopsis>
<syntax>
<channel_snapshot/>
<channel_snapshot prefix="Dest"/>
<parameter name="DialStatus">
<para> The new state of the outbound dial attempt.</para>
<enumlist>
<enum name="RINGING">
<para>The outbound channel is ringing.</para>
</enum>
<enum name="PROCEEDING">
<para>The call to the outbound channel is proceeding.</para>
</enum>
<enum name="PROGRESS">
<para>Progress has been received on the outbound channel.</para>
</enum>
</enumlist>
</parameter>
<parameter name="Forward" required="false">
<para>If the call was forwarded, where the call was
forwarded to.</para>
</parameter>
</syntax>
</managerEventInstance>
</managerEvent>
<managerEvent language="en_US" name="DialEnd">
<managerEventInstance class="EVENT_FLAG_CALL">
<synopsis>Raised when a dial action has completed.</synopsis>
<syntax>
<channel_snapshot/>
<channel_snapshot prefix="Dest"/>
<parameter name="DialStatus">
<para>The result of the dial operation.</para>
<enumlist>
CDRs: fix a variety of dial status problems, h/hangup handler creating CDRs This patch fixes a number of small-ish problems that were noticed when witnessing the records that the FreePBX dialplan produces: (1) Mid-call events (as well as privacy options) have the ability to change the overall state of the Dial operation after the called party answers. This means that publishing the DialEnd event when the called party is premature; we have to wait for the execution of these subroutines to complete before we can signal the overall status of the DialEnd. This patch moves that publication and adds handlers for the mid-call events. (2) The AST_FLAG_OUTGOING channel flag is cleared if an after bridge goto datastore is detected. This flag was preventing CDRs from being recorded for all outbound channels that had a 'continue' option enabled on them by the Dial application. (3) The CDR engine now locks the 'Dial' application as being the CDR application if it detects that the current CDR has entered that app. This is similar to the logic that is done for Parking. In general, if we entered into Dial, then we want that CDR to record the application as such - this prevents pre-dial handlers, mid-call handlers, and other shenaniganry from changing the application value. (4) The CDR engine now checks for the AST_SOFTHANGUP_HANGUP_EXEC in more places to determine if the channel is in hangup logic or dead. In either case, we don't want to record changes in the channel. (5) The default option for "endbeforehexten" has been changed to "yes". In general, you don't want to see CDRs in the 'h' exten or in hangup logic. Since the semantics of that option changed in 12, it made sense to update the default value as well. (6) Finally, because we now have the ability to synchronize on the messages published to the CDR topic, on shutdown the CDR engine will now synchronize to the messages currently in flight. This helps to ensure that all in-flight CDRs are written before shutting down. (closes issue ASTERISK-23164) Reported by: Matt Jordan Review: https://reviewboard.asterisk.org/r/3154 ........ Merged revisions 407084 from http://svn.asterisk.org/svn/asterisk/branches/12 git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@407085 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2014-01-31 23:40:51 +00:00
<enum name="ABORT">
<para>The call was aborted.</para>
</enum>
<enum name="ANSWER">
<para>The caller answered.</para>
</enum>
<enum name="BUSY">
<para>The caller was busy.</para>
</enum>
<enum name="CANCEL">
<para>The caller cancelled the call.</para>
</enum>
<enum name="CHANUNAVAIL">
<para>The requested channel is unavailable.</para>
</enum>
<enum name="CONGESTION">
<para>The called party is congested.</para>
</enum>
<enum name="CONTINUE">
<para>The dial completed, but the caller elected
to continue in the dialplan.</para>
</enum>
<enum name="GOTO">
<para>The dial completed, but the caller jumped to
a dialplan location.</para>
<para>If known, the location the caller is jumping
to will be appended to the result following a
":".</para>
</enum>
<enum name="NOANSWER">
<para>The called party failed to answer.</para>
</enum>
</enumlist>
</parameter>
<parameter name="Forward" required="false">
<para>If the call was forwarded, where the call was
forwarded to.</para>
</parameter>
</syntax>
<see-also>
<ref type="application">Dial</ref>
<ref type="application">Originate</ref>
<ref type="manager">Originate</ref>
<ref type="managerEvent">DialBegin</ref>
</see-also>
</managerEventInstance>
</managerEvent>
<managerEvent language="en_US" name="Hold">
<managerEventInstance class="EVENT_FLAG_CALL">
<synopsis>Raised when a channel goes on hold.</synopsis>
<syntax>
<channel_snapshot/>
<parameter name="MusicClass">
<para>The suggested MusicClass, if provided.</para>
</parameter>
</syntax>
<see-also>
<ref type="managerEvent">Unhold</ref>
</see-also>
</managerEventInstance>
</managerEvent>
<managerEvent language="en_US" name="Unhold">
<managerEventInstance class="EVENT_FLAG_CALL">
<synopsis>Raised when a channel goes off hold.</synopsis>
<syntax>
<channel_snapshot/>
</syntax>
<see-also>
<ref type="managerEvent">Hold</ref>
</see-also>
</managerEventInstance>
</managerEvent>
<managerEvent language="en_US" name="ChanSpyStart">
<managerEventInstance class="EVENT_FLAG_CALL">
<synopsis>Raised when one channel begins spying on another channel.</synopsis>
<syntax>
<channel_snapshot prefix="Spyer"/>
<channel_snapshot prefix="Spyee"/>
</syntax>
<see-also>
<ref type="managerEvent">ChanSpyStop</ref>
<ref type="application">ChanSpy</ref>
</see-also>
</managerEventInstance>
</managerEvent>
<managerEvent language="en_US" name="ChanSpyStop">
<managerEventInstance class="EVENT_FLAG_CALL">
<synopsis>Raised when a channel has stopped spying.</synopsis>
<syntax>
<channel_snapshot prefix="Spyer"/>
<channel_snapshot prefix="Spyee"/>
</syntax>
<see-also>
<ref type="managerEvent">ChanSpyStart</ref>
<ref type="application">ChanSpy</ref>
</see-also>
</managerEventInstance>
</managerEvent>
<managerEvent language="en_US" name="HangupHandlerRun">
<managerEventInstance class="EVENT_FLAG_DIALPLAN">
<synopsis>Raised when a hangup handler is about to be called.</synopsis>
<syntax>
<channel_snapshot/>
<parameter name="Handler">
<para>Hangup handler parameter string passed to the Gosub application.</para>
</parameter>
</syntax>
<see-also>
<ref type="function">CHANNEL</ref>
</see-also>
</managerEventInstance>
</managerEvent>
<managerEvent language="en_US" name="HangupHandlerPop">
<managerEventInstance class="EVENT_FLAG_DIALPLAN">
<synopsis>
Raised when a hangup handler is removed from the handler stack
by the CHANNEL() function.
</synopsis>
<syntax>
<channel_snapshot/>
<xi:include xpointer="xpointer(/docs/managerEvent[@name='HangupHandlerRun']/managerEventInstance/syntax/parameter)" />
</syntax>
<see-also>
<ref type="managerEvent">HangupHandlerPush</ref>
<ref type="function">CHANNEL</ref>
</see-also>
</managerEventInstance>
</managerEvent>
<managerEvent language="en_US" name="HangupHandlerPush">
<managerEventInstance class="EVENT_FLAG_DIALPLAN">
<synopsis>
Raised when a hangup handler is added to the handler stack by
the CHANNEL() function.
</synopsis>
<syntax>
<channel_snapshot/>
<xi:include xpointer="xpointer(/docs/managerEvent[@name='HangupHandlerRun']/managerEventInstance/syntax/parameter)" />
</syntax>
<see-also>
<ref type="managerEvent">HangupHandlerPop</ref>
<ref type="function">CHANNEL</ref>
</see-also>
</managerEventInstance>
</managerEvent>
<managerEvent language="en_US" name="FAXStatus">
<managerEventInstance class="EVENT_FLAG_CALL">
<synopsis>
Raised periodically during a fax transmission.
</synopsis>
<syntax>
<channel_snapshot/>
<parameter name="Operation">
<enumlist>
<enum name="gateway"/>
<enum name="receive"/>
<enum name="send"/>
</enumlist>
</parameter>
<parameter name="Status">
<para>A text message describing the current status of the fax</para>
</parameter>
<xi:include xpointer="xpointer(/docs/managerEvent[@name='ReceiveFAX']/managerEventInstance/syntax/parameter[@name='LocalStationID'])" />
<xi:include xpointer="xpointer(/docs/managerEvent[@name='ReceiveFAX']/managerEventInstance/syntax/parameter[@name='FileName'])" />
</syntax>
</managerEventInstance>
</managerEvent>
<managerEvent language="en_US" name="ReceiveFAX">
<managerEventInstance class="EVENT_FLAG_CALL">
<synopsis>
Raised when a receive fax operation has completed.
</synopsis>
<syntax>
<channel_snapshot/>
<parameter name="LocalStationID">
<para>The value of the <variable>LOCALSTATIONID</variable> channel variable</para>
</parameter>
<parameter name="RemoteStationID">
<para>The value of the <variable>REMOTESTATIONID</variable> channel variable</para>
</parameter>
<parameter name="PagesTransferred">
<para>The number of pages that have been transferred</para>
</parameter>
<parameter name="Resolution">
<para>The negotiated resolution</para>
</parameter>
<parameter name="TransferRate">
<para>The negotiated transfer rate</para>
</parameter>
<parameter name="FileName" multiple="yes">
<para>The files being affected by the fax operation</para>
</parameter>
</syntax>
</managerEventInstance>
</managerEvent>
<managerEvent language="en_US" name="SendFAX">
<managerEventInstance class="EVENT_FLAG_CALL">
<synopsis>
Raised when a send fax operation has completed.
</synopsis>
<syntax>
<channel_snapshot/>
<xi:include xpointer="xpointer(/docs/managerEvent[@name='ReceiveFAX']/managerEventInstance/syntax/parameter)" />
</syntax>
</managerEventInstance>
</managerEvent>
<managerEvent language="en_US" name="MusicOnHoldStart">
<managerEventInstance class="EVENT_FLAG_CALL">
<synopsis>Raised when music on hold has started on a channel.</synopsis>
<syntax>
<channel_snapshot/>
<parameter name="Class">
<para>The class of music being played on the channel</para>
</parameter>
</syntax>
<see-also>
<ref type="managerEvent">MusicOnHoldStop</ref>
<ref type="application">StartMusicOnHold</ref>
<ref type="application">MusicOnHold</ref>
</see-also>
</managerEventInstance>
</managerEvent>
<managerEvent language="en_US" name="MusicOnHoldStop">
<managerEventInstance class="EVENT_FLAG_CALL">
<synopsis>Raised when music on hold has stopped on a channel.</synopsis>
<syntax>
<channel_snapshot/>
</syntax>
<see-also>
<ref type="managerEvent">MusicOnHoldStart</ref>
<ref type="application">StopMusicOnHold</ref>
</see-also>
</managerEventInstance>
</managerEvent>
<managerEvent language="en_US" name="MonitorStart">
<managerEventInstance class="EVENT_FLAG_CALL">
<synopsis>Raised when monitoring has started on a channel.</synopsis>
<syntax>
<channel_snapshot/>
</syntax>
<see-also>
<ref type="managerEvent">MonitorStop</ref>
<ref type="application">Monitor</ref>
<ref type="manager">Monitor</ref>
</see-also>
</managerEventInstance>
</managerEvent>
<managerEvent language="en_US" name="MonitorStop">
<managerEventInstance class="EVENT_FLAG_CALL">
<synopsis>Raised when monitoring has stopped on a channel.</synopsis>
<syntax>
<channel_snapshot/>
</syntax>
<see-also>
<ref type="managerEvent">MonitorStart</ref>
<ref type="application">StopMonitor</ref>
<ref type="manager">StopMonitor</ref>
</see-also>
</managerEventInstance>
</managerEvent>
***/
/*! \brief The \ref stasis subscription returned by the forwarding of the channel topic
* to the manager topic
*/
Multiple revisions 399887,400138,400178,400180-400181 ........ r399887 | dlee | 2013-09-26 10:41:47 -0500 (Thu, 26 Sep 2013) | 1 line Minor performance bump by not allocate manager variable struct if we don't need it ........ r400138 | dlee | 2013-09-30 10:24:00 -0500 (Mon, 30 Sep 2013) | 23 lines Stasis performance improvements This patch addresses several performance problems that were found in the initial performance testing of Asterisk 12. The Stasis dispatch object was allocated as an AO2 object, even though it has a very confined lifecycle. This was replaced with a straight ast_malloc(). The Stasis message router was spending an inordinate amount of time searching hash tables. In this case, most of our routers had 6 or fewer routes in them to begin with. This was replaced with an array that's searched linearly for the route. We more heavily rely on AO2 objects in Asterisk 12, and the memset() in ao2_ref() actually became noticeable on the profile. This was #ifdef'ed to only run when AO2_DEBUG was enabled. After being misled by an erroneous comment in taskprocessor.c during profiling, the wrong comment was removed. Review: https://reviewboard.asterisk.org/r/2873/ ........ r400178 | dlee | 2013-09-30 13:26:27 -0500 (Mon, 30 Sep 2013) | 24 lines Taskprocessor optimization; switch Stasis to use taskprocessors This patch optimizes taskprocessor to use a semaphore for signaling, which the OS can do a better job at managing contention and waiting that we can with a mutex and condition. The taskprocessor execution was also slightly optimized to reduce the number of locks taken. The only observable difference in the taskprocessor implementation is that when the final reference to the taskprocessor goes away, it will execute all tasks to completion instead of discarding the unexecuted tasks. For systems where unnamed semaphores are not supported, a really simple semaphore implementation is provided. (Which gives identical performance as the original taskprocessor implementation). The way we ended up implementing Stasis caused the threadpool to be a burden instead of a boost to performance. This was switched to just use taskprocessors directly for subscriptions. Review: https://reviewboard.asterisk.org/r/2881/ ........ r400180 | dlee | 2013-09-30 13:39:34 -0500 (Mon, 30 Sep 2013) | 28 lines Optimize how Stasis forwards are dispatched This patch optimizes how forwards are dispatched in Stasis. Originally, forwards were dispatched as subscriptions that are invoked on the publishing thread. This did not account for the vast number of forwards we would end up having in the system, and the amount of work it would take to walk though the forward subscriptions. This patch modifies Stasis so that rather than walking the tree of forwards on every dispatch, when forwards and subscriptions are changed, the subscriber list for every topic in the tree is changed. This has a couple of benefits. First, this reduces the workload of dispatching messages. It also reduces contention when dispatching to different topics that happen to forward to the same aggregation topic (as happens with all of the channel, bridge and endpoint topics). Since forwards are no longer subscriptions, the bulk of this patch is simply changing stasis_subscription objects to stasis_forward objects (which, admittedly, I should have done in the first place.) Since this required me to yet again put in a growing array, I finally abstracted that out into a set of ast_vector macros in asterisk/vector.h. Review: https://reviewboard.asterisk.org/r/2883/ ........ r400181 | dlee | 2013-09-30 13:48:57 -0500 (Mon, 30 Sep 2013) | 28 lines Remove dispatch object allocation from Stasis publishing While looking for areas for performance improvement, I realized that an unused feature in Stasis was negatively impacting performance. When a message is sent to a subscriber, a dispatch object is allocated for the dispatch, containing the topic the message was published to, the subscriber the message is being sent to, and the message itself. The topic is actually unused by any subscriber in Asterisk today. And the subscriber is associated with the taskprocessor the message is being dispatched to. First, this patch removes the unused topic parameter from Stasis subscription callbacks. Second, this patch introduces the concept of taskprocessor local data, data that may be set on a taskprocessor and provided along with the data pointer when a task is pushed using the ast_taskprocessor_push_local() call. This allows the task to have both data specific to that taskprocessor, in addition to data specific to that invocation. With those two changes, the dispatch object can be removed completely, and the message is simply refcounted and sent directly to the taskprocessor. Review: https://reviewboard.asterisk.org/r/2884/ ........ Merged revisions 399887,400138,400178,400180-400181 from http://svn.asterisk.org/svn/asterisk/branches/12 git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@400186 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2013-09-30 18:55:27 +00:00
static struct stasis_forward *topic_forwarder;
Move more channel events to Stasis; move res_json.c to main/json.c. This patch started out simply as fixing the bouncing tests introduced in r382685, but required some other changes to give it a decent implementation. To fix the bouncing tests, the UserEvent and Newexten AMI events needed to be refactored to dispatch via Stasis. Dispatching directly to AMI resulted in those events sometimes getting ahead of the associated Newchannel events, which would understandably confuse anyone. I found that instead of creating a zillion different message types and structures associated with them, it would be preferable to define a message type that has a channel snapshot and a blob of structured data with a small bit of additional information. The JSON object model provides a very nice way of representing structured data, so I went with that. * Move JSON support from res_json.c to main/json.c * Made libjansson-dev a required dependency * Added an ast_channel_blob message type, which has a channel snapshot and JSON blob of data. * Changed UserEvent and Newexten events so that they are dispatched via ast_channel_blob messages on the channel's topic. * Got rid of the ast_channel_varset message; used ast_channel_blob instead. * Extracted the manager functions converting Stasis channel events to AMI events into manager_channel.c. (issue ASTERISK-21096) Review: https://reviewboard.asterisk.org/r/2381/ git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@383579 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2013-03-22 14:06:46 +00:00
struct ast_str *ast_manager_build_channel_state_string_prefix(
const struct ast_channel_snapshot *snapshot,
const char *prefix)
Move more channel events to Stasis; move res_json.c to main/json.c. This patch started out simply as fixing the bouncing tests introduced in r382685, but required some other changes to give it a decent implementation. To fix the bouncing tests, the UserEvent and Newexten AMI events needed to be refactored to dispatch via Stasis. Dispatching directly to AMI resulted in those events sometimes getting ahead of the associated Newchannel events, which would understandably confuse anyone. I found that instead of creating a zillion different message types and structures associated with them, it would be preferable to define a message type that has a channel snapshot and a blob of structured data with a small bit of additional information. The JSON object model provides a very nice way of representing structured data, so I went with that. * Move JSON support from res_json.c to main/json.c * Made libjansson-dev a required dependency * Added an ast_channel_blob message type, which has a channel snapshot and JSON blob of data. * Changed UserEvent and Newexten events so that they are dispatched via ast_channel_blob messages on the channel's topic. * Got rid of the ast_channel_varset message; used ast_channel_blob instead. * Extracted the manager functions converting Stasis channel events to AMI events into manager_channel.c. (issue ASTERISK-21096) Review: https://reviewboard.asterisk.org/r/2381/ git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@383579 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2013-03-22 14:06:46 +00:00
{
struct ast_str *out;
char *caller_name;
char *connected_name;
int res;
if (!snapshot || (snapshot->base->tech_properties & AST_CHAN_TP_INTERNAL)) {
Move more channel events to Stasis; move res_json.c to main/json.c. This patch started out simply as fixing the bouncing tests introduced in r382685, but required some other changes to give it a decent implementation. To fix the bouncing tests, the UserEvent and Newexten AMI events needed to be refactored to dispatch via Stasis. Dispatching directly to AMI resulted in those events sometimes getting ahead of the associated Newchannel events, which would understandably confuse anyone. I found that instead of creating a zillion different message types and structures associated with them, it would be preferable to define a message type that has a channel snapshot and a blob of structured data with a small bit of additional information. The JSON object model provides a very nice way of representing structured data, so I went with that. * Move JSON support from res_json.c to main/json.c * Made libjansson-dev a required dependency * Added an ast_channel_blob message type, which has a channel snapshot and JSON blob of data. * Changed UserEvent and Newexten events so that they are dispatched via ast_channel_blob messages on the channel's topic. * Got rid of the ast_channel_varset message; used ast_channel_blob instead. * Extracted the manager functions converting Stasis channel events to AMI events into manager_channel.c. (issue ASTERISK-21096) Review: https://reviewboard.asterisk.org/r/2381/ git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@383579 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2013-03-22 14:06:46 +00:00
return NULL;
}
out = ast_str_create(1024);
if (!out) {
return NULL;
}
caller_name = ast_escape_c_alloc(snapshot->caller->name);
connected_name = ast_escape_c_alloc(snapshot->connected->name);
Move more channel events to Stasis; move res_json.c to main/json.c. This patch started out simply as fixing the bouncing tests introduced in r382685, but required some other changes to give it a decent implementation. To fix the bouncing tests, the UserEvent and Newexten AMI events needed to be refactored to dispatch via Stasis. Dispatching directly to AMI resulted in those events sometimes getting ahead of the associated Newchannel events, which would understandably confuse anyone. I found that instead of creating a zillion different message types and structures associated with them, it would be preferable to define a message type that has a channel snapshot and a blob of structured data with a small bit of additional information. The JSON object model provides a very nice way of representing structured data, so I went with that. * Move JSON support from res_json.c to main/json.c * Made libjansson-dev a required dependency * Added an ast_channel_blob message type, which has a channel snapshot and JSON blob of data. * Changed UserEvent and Newexten events so that they are dispatched via ast_channel_blob messages on the channel's topic. * Got rid of the ast_channel_varset message; used ast_channel_blob instead. * Extracted the manager functions converting Stasis channel events to AMI events into manager_channel.c. (issue ASTERISK-21096) Review: https://reviewboard.asterisk.org/r/2381/ git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@383579 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2013-03-22 14:06:46 +00:00
res = ast_str_set(&out, 0,
"%sChannel: %s\r\n"
"%sChannelState: %u\r\n"
"%sChannelStateDesc: %s\r\n"
"%sCallerIDNum: %s\r\n"
"%sCallerIDName: %s\r\n"
"%sConnectedLineNum: %s\r\n"
"%sConnectedLineName: %s\r\n"
"%sLanguage: %s\r\n"
"%sAccountCode: %s\r\n"
"%sContext: %s\r\n"
"%sExten: %s\r\n"
"%sPriority: %d\r\n"
"%sUniqueid: %s\r\n"
"%sLinkedid: %s\r\n",
prefix, snapshot->base->name,
prefix, snapshot->state,
prefix, ast_state2str(snapshot->state),
prefix, S_OR(snapshot->caller->number, "<unknown>"),
prefix, S_OR(caller_name, "<unknown>"),
prefix, S_OR(snapshot->connected->number, "<unknown>"),
prefix, S_OR(connected_name, "<unknown>"),
prefix, snapshot->base->language,
prefix, snapshot->base->accountcode,
prefix, snapshot->dialplan->context,
prefix, snapshot->dialplan->exten,
prefix, snapshot->dialplan->priority,
prefix, snapshot->base->uniqueid,
prefix, snapshot->peer->linkedid);
Move more channel events to Stasis; move res_json.c to main/json.c. This patch started out simply as fixing the bouncing tests introduced in r382685, but required some other changes to give it a decent implementation. To fix the bouncing tests, the UserEvent and Newexten AMI events needed to be refactored to dispatch via Stasis. Dispatching directly to AMI resulted in those events sometimes getting ahead of the associated Newchannel events, which would understandably confuse anyone. I found that instead of creating a zillion different message types and structures associated with them, it would be preferable to define a message type that has a channel snapshot and a blob of structured data with a small bit of additional information. The JSON object model provides a very nice way of representing structured data, so I went with that. * Move JSON support from res_json.c to main/json.c * Made libjansson-dev a required dependency * Added an ast_channel_blob message type, which has a channel snapshot and JSON blob of data. * Changed UserEvent and Newexten events so that they are dispatched via ast_channel_blob messages on the channel's topic. * Got rid of the ast_channel_varset message; used ast_channel_blob instead. * Extracted the manager functions converting Stasis channel events to AMI events into manager_channel.c. (issue ASTERISK-21096) Review: https://reviewboard.asterisk.org/r/2381/ git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@383579 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2013-03-22 14:06:46 +00:00
ast_free(caller_name);
ast_free(connected_name);
Move more channel events to Stasis; move res_json.c to main/json.c. This patch started out simply as fixing the bouncing tests introduced in r382685, but required some other changes to give it a decent implementation. To fix the bouncing tests, the UserEvent and Newexten AMI events needed to be refactored to dispatch via Stasis. Dispatching directly to AMI resulted in those events sometimes getting ahead of the associated Newchannel events, which would understandably confuse anyone. I found that instead of creating a zillion different message types and structures associated with them, it would be preferable to define a message type that has a channel snapshot and a blob of structured data with a small bit of additional information. The JSON object model provides a very nice way of representing structured data, so I went with that. * Move JSON support from res_json.c to main/json.c * Made libjansson-dev a required dependency * Added an ast_channel_blob message type, which has a channel snapshot and JSON blob of data. * Changed UserEvent and Newexten events so that they are dispatched via ast_channel_blob messages on the channel's topic. * Got rid of the ast_channel_varset message; used ast_channel_blob instead. * Extracted the manager functions converting Stasis channel events to AMI events into manager_channel.c. (issue ASTERISK-21096) Review: https://reviewboard.asterisk.org/r/2381/ git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@383579 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2013-03-22 14:06:46 +00:00
if (!res) {
ast_free(out);
Move more channel events to Stasis; move res_json.c to main/json.c. This patch started out simply as fixing the bouncing tests introduced in r382685, but required some other changes to give it a decent implementation. To fix the bouncing tests, the UserEvent and Newexten AMI events needed to be refactored to dispatch via Stasis. Dispatching directly to AMI resulted in those events sometimes getting ahead of the associated Newchannel events, which would understandably confuse anyone. I found that instead of creating a zillion different message types and structures associated with them, it would be preferable to define a message type that has a channel snapshot and a blob of structured data with a small bit of additional information. The JSON object model provides a very nice way of representing structured data, so I went with that. * Move JSON support from res_json.c to main/json.c * Made libjansson-dev a required dependency * Added an ast_channel_blob message type, which has a channel snapshot and JSON blob of data. * Changed UserEvent and Newexten events so that they are dispatched via ast_channel_blob messages on the channel's topic. * Got rid of the ast_channel_varset message; used ast_channel_blob instead. * Extracted the manager functions converting Stasis channel events to AMI events into manager_channel.c. (issue ASTERISK-21096) Review: https://reviewboard.asterisk.org/r/2381/ git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@383579 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2013-03-22 14:06:46 +00:00
return NULL;
}
if (snapshot->manager_vars) {
struct ast_var_t *var;
char *val;
AST_LIST_TRAVERSE(snapshot->manager_vars, var, entries) {
val = ast_escape_c_alloc(var->value);
ast_str_append(&out, 0, "%sChanVariable: %s=%s\r\n",
prefix,
var->name, S_OR(val, ""));
ast_free(val);
}
}
Move more channel events to Stasis; move res_json.c to main/json.c. This patch started out simply as fixing the bouncing tests introduced in r382685, but required some other changes to give it a decent implementation. To fix the bouncing tests, the UserEvent and Newexten AMI events needed to be refactored to dispatch via Stasis. Dispatching directly to AMI resulted in those events sometimes getting ahead of the associated Newchannel events, which would understandably confuse anyone. I found that instead of creating a zillion different message types and structures associated with them, it would be preferable to define a message type that has a channel snapshot and a blob of structured data with a small bit of additional information. The JSON object model provides a very nice way of representing structured data, so I went with that. * Move JSON support from res_json.c to main/json.c * Made libjansson-dev a required dependency * Added an ast_channel_blob message type, which has a channel snapshot and JSON blob of data. * Changed UserEvent and Newexten events so that they are dispatched via ast_channel_blob messages on the channel's topic. * Got rid of the ast_channel_varset message; used ast_channel_blob instead. * Extracted the manager functions converting Stasis channel events to AMI events into manager_channel.c. (issue ASTERISK-21096) Review: https://reviewboard.asterisk.org/r/2381/ git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@383579 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2013-03-22 14:06:46 +00:00
return out;
}
struct ast_str *ast_manager_build_channel_state_string(
const struct ast_channel_snapshot *snapshot)
{
return ast_manager_build_channel_state_string_prefix(snapshot, "");
}
/*! \brief Typedef for callbacks that get called on channel snapshot updates */
typedef struct ast_manager_event_blob *(*channel_snapshot_monitor)(
struct ast_channel_snapshot *old_snapshot,
struct ast_channel_snapshot *new_snapshot);
/*! \brief Handle channel state changes */
static struct ast_manager_event_blob *channel_state_change(
struct ast_channel_snapshot *old_snapshot,
struct ast_channel_snapshot *new_snapshot)
{
int is_hungup, was_hungup;
Move more channel events to Stasis; move res_json.c to main/json.c. This patch started out simply as fixing the bouncing tests introduced in r382685, but required some other changes to give it a decent implementation. To fix the bouncing tests, the UserEvent and Newexten AMI events needed to be refactored to dispatch via Stasis. Dispatching directly to AMI resulted in those events sometimes getting ahead of the associated Newchannel events, which would understandably confuse anyone. I found that instead of creating a zillion different message types and structures associated with them, it would be preferable to define a message type that has a channel snapshot and a blob of structured data with a small bit of additional information. The JSON object model provides a very nice way of representing structured data, so I went with that. * Move JSON support from res_json.c to main/json.c * Made libjansson-dev a required dependency * Added an ast_channel_blob message type, which has a channel snapshot and JSON blob of data. * Changed UserEvent and Newexten events so that they are dispatched via ast_channel_blob messages on the channel's topic. * Got rid of the ast_channel_varset message; used ast_channel_blob instead. * Extracted the manager functions converting Stasis channel events to AMI events into manager_channel.c. (issue ASTERISK-21096) Review: https://reviewboard.asterisk.org/r/2381/ git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@383579 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2013-03-22 14:06:46 +00:00
/* The Newchannel, Newstate and Hangup events are closely related, in
* in that they are mutually exclusive, basically different flavors
* of a new channel state event.
*/
Move more channel events to Stasis; move res_json.c to main/json.c. This patch started out simply as fixing the bouncing tests introduced in r382685, but required some other changes to give it a decent implementation. To fix the bouncing tests, the UserEvent and Newexten AMI events needed to be refactored to dispatch via Stasis. Dispatching directly to AMI resulted in those events sometimes getting ahead of the associated Newchannel events, which would understandably confuse anyone. I found that instead of creating a zillion different message types and structures associated with them, it would be preferable to define a message type that has a channel snapshot and a blob of structured data with a small bit of additional information. The JSON object model provides a very nice way of representing structured data, so I went with that. * Move JSON support from res_json.c to main/json.c * Made libjansson-dev a required dependency * Added an ast_channel_blob message type, which has a channel snapshot and JSON blob of data. * Changed UserEvent and Newexten events so that they are dispatched via ast_channel_blob messages on the channel's topic. * Got rid of the ast_channel_varset message; used ast_channel_blob instead. * Extracted the manager functions converting Stasis channel events to AMI events into manager_channel.c. (issue ASTERISK-21096) Review: https://reviewboard.asterisk.org/r/2381/ git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@383579 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2013-03-22 14:06:46 +00:00
if (!old_snapshot) {
return ast_manager_event_blob_create(
EVENT_FLAG_CALL, "Newchannel", NO_EXTRA_FIELDS);
Move more channel events to Stasis; move res_json.c to main/json.c. This patch started out simply as fixing the bouncing tests introduced in r382685, but required some other changes to give it a decent implementation. To fix the bouncing tests, the UserEvent and Newexten AMI events needed to be refactored to dispatch via Stasis. Dispatching directly to AMI resulted in those events sometimes getting ahead of the associated Newchannel events, which would understandably confuse anyone. I found that instead of creating a zillion different message types and structures associated with them, it would be preferable to define a message type that has a channel snapshot and a blob of structured data with a small bit of additional information. The JSON object model provides a very nice way of representing structured data, so I went with that. * Move JSON support from res_json.c to main/json.c * Made libjansson-dev a required dependency * Added an ast_channel_blob message type, which has a channel snapshot and JSON blob of data. * Changed UserEvent and Newexten events so that they are dispatched via ast_channel_blob messages on the channel's topic. * Got rid of the ast_channel_varset message; used ast_channel_blob instead. * Extracted the manager functions converting Stasis channel events to AMI events into manager_channel.c. (issue ASTERISK-21096) Review: https://reviewboard.asterisk.org/r/2381/ git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@383579 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2013-03-22 14:06:46 +00:00
}
was_hungup = ast_test_flag(&old_snapshot->flags, AST_FLAG_DEAD) ? 1 : 0;
is_hungup = ast_test_flag(&new_snapshot->flags, AST_FLAG_DEAD) ? 1 : 0;
Move more channel events to Stasis; move res_json.c to main/json.c. This patch started out simply as fixing the bouncing tests introduced in r382685, but required some other changes to give it a decent implementation. To fix the bouncing tests, the UserEvent and Newexten AMI events needed to be refactored to dispatch via Stasis. Dispatching directly to AMI resulted in those events sometimes getting ahead of the associated Newchannel events, which would understandably confuse anyone. I found that instead of creating a zillion different message types and structures associated with them, it would be preferable to define a message type that has a channel snapshot and a blob of structured data with a small bit of additional information. The JSON object model provides a very nice way of representing structured data, so I went with that. * Move JSON support from res_json.c to main/json.c * Made libjansson-dev a required dependency * Added an ast_channel_blob message type, which has a channel snapshot and JSON blob of data. * Changed UserEvent and Newexten events so that they are dispatched via ast_channel_blob messages on the channel's topic. * Got rid of the ast_channel_varset message; used ast_channel_blob instead. * Extracted the manager functions converting Stasis channel events to AMI events into manager_channel.c. (issue ASTERISK-21096) Review: https://reviewboard.asterisk.org/r/2381/ git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@383579 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2013-03-22 14:06:46 +00:00
if (!was_hungup && is_hungup) {
return ast_manager_event_blob_create(
EVENT_FLAG_CALL, "Hangup",
"Cause: %d\r\n"
"Cause-txt: %s\r\n",
new_snapshot->hangup->cause,
ast_cause2str(new_snapshot->hangup->cause));
Move more channel events to Stasis; move res_json.c to main/json.c. This patch started out simply as fixing the bouncing tests introduced in r382685, but required some other changes to give it a decent implementation. To fix the bouncing tests, the UserEvent and Newexten AMI events needed to be refactored to dispatch via Stasis. Dispatching directly to AMI resulted in those events sometimes getting ahead of the associated Newchannel events, which would understandably confuse anyone. I found that instead of creating a zillion different message types and structures associated with them, it would be preferable to define a message type that has a channel snapshot and a blob of structured data with a small bit of additional information. The JSON object model provides a very nice way of representing structured data, so I went with that. * Move JSON support from res_json.c to main/json.c * Made libjansson-dev a required dependency * Added an ast_channel_blob message type, which has a channel snapshot and JSON blob of data. * Changed UserEvent and Newexten events so that they are dispatched via ast_channel_blob messages on the channel's topic. * Got rid of the ast_channel_varset message; used ast_channel_blob instead. * Extracted the manager functions converting Stasis channel events to AMI events into manager_channel.c. (issue ASTERISK-21096) Review: https://reviewboard.asterisk.org/r/2381/ git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@383579 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2013-03-22 14:06:46 +00:00
}
if (old_snapshot->state != new_snapshot->state) {
return ast_manager_event_blob_create(
EVENT_FLAG_CALL, "Newstate", NO_EXTRA_FIELDS);
}
/* No event */
return NULL;
}
static struct ast_manager_event_blob *channel_newexten(
struct ast_channel_snapshot *old_snapshot,
struct ast_channel_snapshot *new_snapshot)
{
/* Empty application is not valid for a Newexten event */
if (ast_strlen_zero(new_snapshot->dialplan->appl)) {
return NULL;
}
/* Ignore any updates if we're hungup */
if (ast_test_flag(&new_snapshot->flags, AST_FLAG_DEAD)) {
return NULL;
}
/* Ignore updates if the CEP is unchanged */
if (old_snapshot && ast_channel_snapshot_cep_equal(old_snapshot, new_snapshot)) {
return NULL;
}
/* DEPRECATED: Extension field deprecated in 12; remove in 14 */
return ast_manager_event_blob_create(
EVENT_FLAG_DIALPLAN, "Newexten",
"Extension: %s\r\n"
"Application: %s\r\n"
"AppData: %s\r\n",
new_snapshot->dialplan->exten,
new_snapshot->dialplan->appl,
new_snapshot->dialplan->data);
}
static struct ast_manager_event_blob *channel_new_callerid(
struct ast_channel_snapshot *old_snapshot,
struct ast_channel_snapshot *new_snapshot)
{
struct ast_manager_event_blob *res;
char *callerid;
/* No NewCallerid event on first channel snapshot */
if (!old_snapshot) {
return NULL;
}
if (ast_channel_snapshot_caller_id_equal(old_snapshot, new_snapshot)) {
return NULL;
Move more channel events to Stasis; move res_json.c to main/json.c. This patch started out simply as fixing the bouncing tests introduced in r382685, but required some other changes to give it a decent implementation. To fix the bouncing tests, the UserEvent and Newexten AMI events needed to be refactored to dispatch via Stasis. Dispatching directly to AMI resulted in those events sometimes getting ahead of the associated Newchannel events, which would understandably confuse anyone. I found that instead of creating a zillion different message types and structures associated with them, it would be preferable to define a message type that has a channel snapshot and a blob of structured data with a small bit of additional information. The JSON object model provides a very nice way of representing structured data, so I went with that. * Move JSON support from res_json.c to main/json.c * Made libjansson-dev a required dependency * Added an ast_channel_blob message type, which has a channel snapshot and JSON blob of data. * Changed UserEvent and Newexten events so that they are dispatched via ast_channel_blob messages on the channel's topic. * Got rid of the ast_channel_varset message; used ast_channel_blob instead. * Extracted the manager functions converting Stasis channel events to AMI events into manager_channel.c. (issue ASTERISK-21096) Review: https://reviewboard.asterisk.org/r/2381/ git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@383579 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2013-03-22 14:06:46 +00:00
}
if (!(callerid = ast_escape_c_alloc(
ast_describe_caller_presentation(new_snapshot->caller->pres)))) {
return NULL;
}
res = ast_manager_event_blob_create(
EVENT_FLAG_CALL, "NewCallerid",
"CID-CallingPres: %d (%s)\r\n",
new_snapshot->caller->pres,
callerid);
ast_free(callerid);
return res;
}
static struct ast_manager_event_blob *channel_new_connected_line(
struct ast_channel_snapshot *old_snapshot,
struct ast_channel_snapshot *new_snapshot)
{
/* No NewConnectedLine event on first channel snapshot */
if (!old_snapshot) {
return NULL;
}
if (ast_channel_snapshot_connected_line_equal(old_snapshot, new_snapshot)) {
return NULL;
}
return ast_manager_event_blob_create(
EVENT_FLAG_CALL, "NewConnectedLine", "%s", "");
}
Update Asterisk's CDRs for the new bridging framework This patch is the initial push to update Asterisk's CDR engine for the new bridging framework. This patch guts the existing CDR engine and builds the new on top of messages coming across Stasis. As changes in channel state and bridge state are detected, CDRs are built and dispatched accordingly. This fundamentally changes CDRs in a few ways. (1) CDRs are now *very* reflective of the actual state of channels and bridges. This means CDRs track well with what an actual channel is doing - which is useful in transfer scenarios (which were previously difficult to pin down). It does, however, mean that CDRs cannot be 'fooled'. Previous behavior in Asterisk allowed for CDR applications, channels, and other properties to be spoofed in parts of the code - this no longer works. (2) CDRs have defined behavior in multi-party scenarios. This behavior will not be what everyone wants, but it is a defined behavior and as such, it is predictable. (3) The CDR manipulation functions and applications have been overhauled. Major changes have been made to ResetCDR and ForkCDR in particular. Many of the options for these two applications no longer made any sense with the new framework and the (slightly) more immutable nature of CDRs. There are a plethora of other changes. For a full description of CDR behavior, see the CDR specification on the Asterisk wiki. (closes issue ASTERISK-21196) Review: https://reviewboard.asterisk.org/r/2486/ git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@391947 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2013-06-17 03:00:38 +00:00
static struct ast_manager_event_blob *channel_new_accountcode(
struct ast_channel_snapshot *old_snapshot,
struct ast_channel_snapshot *new_snapshot)
{
if (!old_snapshot) {
Update Asterisk's CDRs for the new bridging framework This patch is the initial push to update Asterisk's CDR engine for the new bridging framework. This patch guts the existing CDR engine and builds the new on top of messages coming across Stasis. As changes in channel state and bridge state are detected, CDRs are built and dispatched accordingly. This fundamentally changes CDRs in a few ways. (1) CDRs are now *very* reflective of the actual state of channels and bridges. This means CDRs track well with what an actual channel is doing - which is useful in transfer scenarios (which were previously difficult to pin down). It does, however, mean that CDRs cannot be 'fooled'. Previous behavior in Asterisk allowed for CDR applications, channels, and other properties to be spoofed in parts of the code - this no longer works. (2) CDRs have defined behavior in multi-party scenarios. This behavior will not be what everyone wants, but it is a defined behavior and as such, it is predictable. (3) The CDR manipulation functions and applications have been overhauled. Major changes have been made to ResetCDR and ForkCDR in particular. Many of the options for these two applications no longer made any sense with the new framework and the (slightly) more immutable nature of CDRs. There are a plethora of other changes. For a full description of CDR behavior, see the CDR specification on the Asterisk wiki. (closes issue ASTERISK-21196) Review: https://reviewboard.asterisk.org/r/2486/ git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@391947 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2013-06-17 03:00:38 +00:00
return NULL;
}
if (!strcmp(old_snapshot->base->accountcode, new_snapshot->base->accountcode)) {
Update Asterisk's CDRs for the new bridging framework This patch is the initial push to update Asterisk's CDR engine for the new bridging framework. This patch guts the existing CDR engine and builds the new on top of messages coming across Stasis. As changes in channel state and bridge state are detected, CDRs are built and dispatched accordingly. This fundamentally changes CDRs in a few ways. (1) CDRs are now *very* reflective of the actual state of channels and bridges. This means CDRs track well with what an actual channel is doing - which is useful in transfer scenarios (which were previously difficult to pin down). It does, however, mean that CDRs cannot be 'fooled'. Previous behavior in Asterisk allowed for CDR applications, channels, and other properties to be spoofed in parts of the code - this no longer works. (2) CDRs have defined behavior in multi-party scenarios. This behavior will not be what everyone wants, but it is a defined behavior and as such, it is predictable. (3) The CDR manipulation functions and applications have been overhauled. Major changes have been made to ResetCDR and ForkCDR in particular. Many of the options for these two applications no longer made any sense with the new framework and the (slightly) more immutable nature of CDRs. There are a plethora of other changes. For a full description of CDR behavior, see the CDR specification on the Asterisk wiki. (closes issue ASTERISK-21196) Review: https://reviewboard.asterisk.org/r/2486/ git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@391947 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2013-06-17 03:00:38 +00:00
return NULL;
}
return ast_manager_event_blob_create(
EVENT_FLAG_CALL, "NewAccountCode",
"OldAccountCode: %s\r\n", old_snapshot->base->accountcode);
Update Asterisk's CDRs for the new bridging framework This patch is the initial push to update Asterisk's CDR engine for the new bridging framework. This patch guts the existing CDR engine and builds the new on top of messages coming across Stasis. As changes in channel state and bridge state are detected, CDRs are built and dispatched accordingly. This fundamentally changes CDRs in a few ways. (1) CDRs are now *very* reflective of the actual state of channels and bridges. This means CDRs track well with what an actual channel is doing - which is useful in transfer scenarios (which were previously difficult to pin down). It does, however, mean that CDRs cannot be 'fooled'. Previous behavior in Asterisk allowed for CDR applications, channels, and other properties to be spoofed in parts of the code - this no longer works. (2) CDRs have defined behavior in multi-party scenarios. This behavior will not be what everyone wants, but it is a defined behavior and as such, it is predictable. (3) The CDR manipulation functions and applications have been overhauled. Major changes have been made to ResetCDR and ForkCDR in particular. Many of the options for these two applications no longer made any sense with the new framework and the (slightly) more immutable nature of CDRs. There are a plethora of other changes. For a full description of CDR behavior, see the CDR specification on the Asterisk wiki. (closes issue ASTERISK-21196) Review: https://reviewboard.asterisk.org/r/2486/ git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@391947 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2013-06-17 03:00:38 +00:00
}
channel_snapshot_monitor channel_monitors[] = {
channel_state_change,
channel_newexten,
Update Asterisk's CDRs for the new bridging framework This patch is the initial push to update Asterisk's CDR engine for the new bridging framework. This patch guts the existing CDR engine and builds the new on top of messages coming across Stasis. As changes in channel state and bridge state are detected, CDRs are built and dispatched accordingly. This fundamentally changes CDRs in a few ways. (1) CDRs are now *very* reflective of the actual state of channels and bridges. This means CDRs track well with what an actual channel is doing - which is useful in transfer scenarios (which were previously difficult to pin down). It does, however, mean that CDRs cannot be 'fooled'. Previous behavior in Asterisk allowed for CDR applications, channels, and other properties to be spoofed in parts of the code - this no longer works. (2) CDRs have defined behavior in multi-party scenarios. This behavior will not be what everyone wants, but it is a defined behavior and as such, it is predictable. (3) The CDR manipulation functions and applications have been overhauled. Major changes have been made to ResetCDR and ForkCDR in particular. Many of the options for these two applications no longer made any sense with the new framework and the (slightly) more immutable nature of CDRs. There are a plethora of other changes. For a full description of CDR behavior, see the CDR specification on the Asterisk wiki. (closes issue ASTERISK-21196) Review: https://reviewboard.asterisk.org/r/2486/ git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@391947 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2013-06-17 03:00:38 +00:00
channel_new_callerid,
channel_new_accountcode,
channel_new_connected_line,
};
static void channel_snapshot_update(void *data, struct stasis_subscription *sub,
struct stasis_message *message)
{
RAII_VAR(struct ast_str *, channel_event_string, NULL, ast_free);
struct ast_channel_snapshot_update *update;
size_t i;
update = stasis_message_data(message);
for (i = 0; i < ARRAY_LEN(channel_monitors); ++i) {
RAII_VAR(struct ast_manager_event_blob *, ev, NULL, ao2_cleanup);
ev = channel_monitors[i](update->old_snapshot, update->new_snapshot);
if (!ev) {
continue;
}
/* If we haven't already, build the channel event string */
if (!channel_event_string) {
channel_event_string =
ast_manager_build_channel_state_string(update->new_snapshot);
if (!channel_event_string) {
return;
}
}
manager_event(ev->event_flags, ev->manager_event, "%s%s",
ast_str_buffer(channel_event_string),
ev->extra_fields);
Move more channel events to Stasis; move res_json.c to main/json.c. This patch started out simply as fixing the bouncing tests introduced in r382685, but required some other changes to give it a decent implementation. To fix the bouncing tests, the UserEvent and Newexten AMI events needed to be refactored to dispatch via Stasis. Dispatching directly to AMI resulted in those events sometimes getting ahead of the associated Newchannel events, which would understandably confuse anyone. I found that instead of creating a zillion different message types and structures associated with them, it would be preferable to define a message type that has a channel snapshot and a blob of structured data with a small bit of additional information. The JSON object model provides a very nice way of representing structured data, so I went with that. * Move JSON support from res_json.c to main/json.c * Made libjansson-dev a required dependency * Added an ast_channel_blob message type, which has a channel snapshot and JSON blob of data. * Changed UserEvent and Newexten events so that they are dispatched via ast_channel_blob messages on the channel's topic. * Got rid of the ast_channel_varset message; used ast_channel_blob instead. * Extracted the manager functions converting Stasis channel events to AMI events into manager_channel.c. (issue ASTERISK-21096) Review: https://reviewboard.asterisk.org/r/2381/ git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@383579 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2013-03-22 14:06:46 +00:00
}
}
static void publish_basic_channel_event(const char *event, int class, struct ast_channel_snapshot *snapshot)
{
RAII_VAR(struct ast_str *, channel_event_string, NULL, ast_free);
channel_event_string = ast_manager_build_channel_state_string(snapshot);
if (!channel_event_string) {
return;
}
manager_event(class, event,
"%s",
ast_str_buffer(channel_event_string));
}
static void channel_hangup_request_cb(void *data,
Multiple revisions 399887,400138,400178,400180-400181 ........ r399887 | dlee | 2013-09-26 10:41:47 -0500 (Thu, 26 Sep 2013) | 1 line Minor performance bump by not allocate manager variable struct if we don't need it ........ r400138 | dlee | 2013-09-30 10:24:00 -0500 (Mon, 30 Sep 2013) | 23 lines Stasis performance improvements This patch addresses several performance problems that were found in the initial performance testing of Asterisk 12. The Stasis dispatch object was allocated as an AO2 object, even though it has a very confined lifecycle. This was replaced with a straight ast_malloc(). The Stasis message router was spending an inordinate amount of time searching hash tables. In this case, most of our routers had 6 or fewer routes in them to begin with. This was replaced with an array that's searched linearly for the route. We more heavily rely on AO2 objects in Asterisk 12, and the memset() in ao2_ref() actually became noticeable on the profile. This was #ifdef'ed to only run when AO2_DEBUG was enabled. After being misled by an erroneous comment in taskprocessor.c during profiling, the wrong comment was removed. Review: https://reviewboard.asterisk.org/r/2873/ ........ r400178 | dlee | 2013-09-30 13:26:27 -0500 (Mon, 30 Sep 2013) | 24 lines Taskprocessor optimization; switch Stasis to use taskprocessors This patch optimizes taskprocessor to use a semaphore for signaling, which the OS can do a better job at managing contention and waiting that we can with a mutex and condition. The taskprocessor execution was also slightly optimized to reduce the number of locks taken. The only observable difference in the taskprocessor implementation is that when the final reference to the taskprocessor goes away, it will execute all tasks to completion instead of discarding the unexecuted tasks. For systems where unnamed semaphores are not supported, a really simple semaphore implementation is provided. (Which gives identical performance as the original taskprocessor implementation). The way we ended up implementing Stasis caused the threadpool to be a burden instead of a boost to performance. This was switched to just use taskprocessors directly for subscriptions. Review: https://reviewboard.asterisk.org/r/2881/ ........ r400180 | dlee | 2013-09-30 13:39:34 -0500 (Mon, 30 Sep 2013) | 28 lines Optimize how Stasis forwards are dispatched This patch optimizes how forwards are dispatched in Stasis. Originally, forwards were dispatched as subscriptions that are invoked on the publishing thread. This did not account for the vast number of forwards we would end up having in the system, and the amount of work it would take to walk though the forward subscriptions. This patch modifies Stasis so that rather than walking the tree of forwards on every dispatch, when forwards and subscriptions are changed, the subscriber list for every topic in the tree is changed. This has a couple of benefits. First, this reduces the workload of dispatching messages. It also reduces contention when dispatching to different topics that happen to forward to the same aggregation topic (as happens with all of the channel, bridge and endpoint topics). Since forwards are no longer subscriptions, the bulk of this patch is simply changing stasis_subscription objects to stasis_forward objects (which, admittedly, I should have done in the first place.) Since this required me to yet again put in a growing array, I finally abstracted that out into a set of ast_vector macros in asterisk/vector.h. Review: https://reviewboard.asterisk.org/r/2883/ ........ r400181 | dlee | 2013-09-30 13:48:57 -0500 (Mon, 30 Sep 2013) | 28 lines Remove dispatch object allocation from Stasis publishing While looking for areas for performance improvement, I realized that an unused feature in Stasis was negatively impacting performance. When a message is sent to a subscriber, a dispatch object is allocated for the dispatch, containing the topic the message was published to, the subscriber the message is being sent to, and the message itself. The topic is actually unused by any subscriber in Asterisk today. And the subscriber is associated with the taskprocessor the message is being dispatched to. First, this patch removes the unused topic parameter from Stasis subscription callbacks. Second, this patch introduces the concept of taskprocessor local data, data that may be set on a taskprocessor and provided along with the data pointer when a task is pushed using the ast_taskprocessor_push_local() call. This allows the task to have both data specific to that taskprocessor, in addition to data specific to that invocation. With those two changes, the dispatch object can be removed completely, and the message is simply refcounted and sent directly to the taskprocessor. Review: https://reviewboard.asterisk.org/r/2884/ ........ Merged revisions 399887,400138,400178,400180-400181 from http://svn.asterisk.org/svn/asterisk/branches/12 git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@400186 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2013-09-30 18:55:27 +00:00
struct stasis_subscription *sub,
struct stasis_message *message)
{
struct ast_channel_blob *obj = stasis_message_data(message);
struct ast_str *extra;
struct ast_str *channel_event_string;
struct ast_json *cause;
int is_soft;
char *manager_event = "HangupRequest";
if (!obj->snapshot) {
/* No snapshot? Likely an earlier allocation failure creating it. */
return;
}
extra = ast_str_create(20);
if (!extra) {
return;
}
channel_event_string = ast_manager_build_channel_state_string(obj->snapshot);
if (!channel_event_string) {
ast_free(extra);
return;
}
cause = ast_json_object_get(obj->blob, "cause");
if (cause) {
ast_str_append(&extra, 0,
"Cause: %jd\r\n",
ast_json_integer_get(cause));
}
is_soft = ast_json_is_true(ast_json_object_get(obj->blob, "soft"));
if (is_soft) {
manager_event = "SoftHangupRequest";
}
manager_event(EVENT_FLAG_CALL, manager_event,
"%s%s",
ast_str_buffer(channel_event_string),
ast_str_buffer(extra));
ast_free(channel_event_string);
ast_free(extra);
}
static void channel_chanspy_stop_cb(void *data, struct stasis_subscription *sub,
Multiple revisions 399887,400138,400178,400180-400181 ........ r399887 | dlee | 2013-09-26 10:41:47 -0500 (Thu, 26 Sep 2013) | 1 line Minor performance bump by not allocate manager variable struct if we don't need it ........ r400138 | dlee | 2013-09-30 10:24:00 -0500 (Mon, 30 Sep 2013) | 23 lines Stasis performance improvements This patch addresses several performance problems that were found in the initial performance testing of Asterisk 12. The Stasis dispatch object was allocated as an AO2 object, even though it has a very confined lifecycle. This was replaced with a straight ast_malloc(). The Stasis message router was spending an inordinate amount of time searching hash tables. In this case, most of our routers had 6 or fewer routes in them to begin with. This was replaced with an array that's searched linearly for the route. We more heavily rely on AO2 objects in Asterisk 12, and the memset() in ao2_ref() actually became noticeable on the profile. This was #ifdef'ed to only run when AO2_DEBUG was enabled. After being misled by an erroneous comment in taskprocessor.c during profiling, the wrong comment was removed. Review: https://reviewboard.asterisk.org/r/2873/ ........ r400178 | dlee | 2013-09-30 13:26:27 -0500 (Mon, 30 Sep 2013) | 24 lines Taskprocessor optimization; switch Stasis to use taskprocessors This patch optimizes taskprocessor to use a semaphore for signaling, which the OS can do a better job at managing contention and waiting that we can with a mutex and condition. The taskprocessor execution was also slightly optimized to reduce the number of locks taken. The only observable difference in the taskprocessor implementation is that when the final reference to the taskprocessor goes away, it will execute all tasks to completion instead of discarding the unexecuted tasks. For systems where unnamed semaphores are not supported, a really simple semaphore implementation is provided. (Which gives identical performance as the original taskprocessor implementation). The way we ended up implementing Stasis caused the threadpool to be a burden instead of a boost to performance. This was switched to just use taskprocessors directly for subscriptions. Review: https://reviewboard.asterisk.org/r/2881/ ........ r400180 | dlee | 2013-09-30 13:39:34 -0500 (Mon, 30 Sep 2013) | 28 lines Optimize how Stasis forwards are dispatched This patch optimizes how forwards are dispatched in Stasis. Originally, forwards were dispatched as subscriptions that are invoked on the publishing thread. This did not account for the vast number of forwards we would end up having in the system, and the amount of work it would take to walk though the forward subscriptions. This patch modifies Stasis so that rather than walking the tree of forwards on every dispatch, when forwards and subscriptions are changed, the subscriber list for every topic in the tree is changed. This has a couple of benefits. First, this reduces the workload of dispatching messages. It also reduces contention when dispatching to different topics that happen to forward to the same aggregation topic (as happens with all of the channel, bridge and endpoint topics). Since forwards are no longer subscriptions, the bulk of this patch is simply changing stasis_subscription objects to stasis_forward objects (which, admittedly, I should have done in the first place.) Since this required me to yet again put in a growing array, I finally abstracted that out into a set of ast_vector macros in asterisk/vector.h. Review: https://reviewboard.asterisk.org/r/2883/ ........ r400181 | dlee | 2013-09-30 13:48:57 -0500 (Mon, 30 Sep 2013) | 28 lines Remove dispatch object allocation from Stasis publishing While looking for areas for performance improvement, I realized that an unused feature in Stasis was negatively impacting performance. When a message is sent to a subscriber, a dispatch object is allocated for the dispatch, containing the topic the message was published to, the subscriber the message is being sent to, and the message itself. The topic is actually unused by any subscriber in Asterisk today. And the subscriber is associated with the taskprocessor the message is being dispatched to. First, this patch removes the unused topic parameter from Stasis subscription callbacks. Second, this patch introduces the concept of taskprocessor local data, data that may be set on a taskprocessor and provided along with the data pointer when a task is pushed using the ast_taskprocessor_push_local() call. This allows the task to have both data specific to that taskprocessor, in addition to data specific to that invocation. With those two changes, the dispatch object can be removed completely, and the message is simply refcounted and sent directly to the taskprocessor. Review: https://reviewboard.asterisk.org/r/2884/ ........ Merged revisions 399887,400138,400178,400180-400181 from http://svn.asterisk.org/svn/asterisk/branches/12 git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@400186 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2013-09-30 18:55:27 +00:00
struct stasis_message *message)
{
RAII_VAR(struct ast_str *, spyer_channel_string, NULL, ast_free);
RAII_VAR(struct ast_str *, spyee_channel_string, NULL, ast_free);
struct ast_channel_snapshot *spyer;
struct ast_channel_snapshot *spyee;
const char *spyee_info = "";
struct ast_multi_channel_blob *payload = stasis_message_data(message);
spyer = ast_multi_channel_blob_get_channel(payload, "spyer_channel");
if (!spyer) {
ast_log(AST_LOG_WARNING, "Received ChanSpy Stop event with no spyer channel!\n");
return;
}
spyer_channel_string = ast_manager_build_channel_state_string_prefix(spyer, "Spyer");
if (!spyer_channel_string) {
return;
}
spyee = ast_multi_channel_blob_get_channel(payload, "spyee_channel");
if (spyee) {
spyee_channel_string = ast_manager_build_channel_state_string_prefix(spyee, "Spyee");
if (spyee_channel_string) {
spyee_info = ast_str_buffer(spyee_channel_string);
}
}
manager_event(EVENT_FLAG_CALL, "ChanSpyStop",
"%s%s",
ast_str_buffer(spyer_channel_string),
spyee_info);
}
static void channel_chanspy_start_cb(void *data, struct stasis_subscription *sub,
Multiple revisions 399887,400138,400178,400180-400181 ........ r399887 | dlee | 2013-09-26 10:41:47 -0500 (Thu, 26 Sep 2013) | 1 line Minor performance bump by not allocate manager variable struct if we don't need it ........ r400138 | dlee | 2013-09-30 10:24:00 -0500 (Mon, 30 Sep 2013) | 23 lines Stasis performance improvements This patch addresses several performance problems that were found in the initial performance testing of Asterisk 12. The Stasis dispatch object was allocated as an AO2 object, even though it has a very confined lifecycle. This was replaced with a straight ast_malloc(). The Stasis message router was spending an inordinate amount of time searching hash tables. In this case, most of our routers had 6 or fewer routes in them to begin with. This was replaced with an array that's searched linearly for the route. We more heavily rely on AO2 objects in Asterisk 12, and the memset() in ao2_ref() actually became noticeable on the profile. This was #ifdef'ed to only run when AO2_DEBUG was enabled. After being misled by an erroneous comment in taskprocessor.c during profiling, the wrong comment was removed. Review: https://reviewboard.asterisk.org/r/2873/ ........ r400178 | dlee | 2013-09-30 13:26:27 -0500 (Mon, 30 Sep 2013) | 24 lines Taskprocessor optimization; switch Stasis to use taskprocessors This patch optimizes taskprocessor to use a semaphore for signaling, which the OS can do a better job at managing contention and waiting that we can with a mutex and condition. The taskprocessor execution was also slightly optimized to reduce the number of locks taken. The only observable difference in the taskprocessor implementation is that when the final reference to the taskprocessor goes away, it will execute all tasks to completion instead of discarding the unexecuted tasks. For systems where unnamed semaphores are not supported, a really simple semaphore implementation is provided. (Which gives identical performance as the original taskprocessor implementation). The way we ended up implementing Stasis caused the threadpool to be a burden instead of a boost to performance. This was switched to just use taskprocessors directly for subscriptions. Review: https://reviewboard.asterisk.org/r/2881/ ........ r400180 | dlee | 2013-09-30 13:39:34 -0500 (Mon, 30 Sep 2013) | 28 lines Optimize how Stasis forwards are dispatched This patch optimizes how forwards are dispatched in Stasis. Originally, forwards were dispatched as subscriptions that are invoked on the publishing thread. This did not account for the vast number of forwards we would end up having in the system, and the amount of work it would take to walk though the forward subscriptions. This patch modifies Stasis so that rather than walking the tree of forwards on every dispatch, when forwards and subscriptions are changed, the subscriber list for every topic in the tree is changed. This has a couple of benefits. First, this reduces the workload of dispatching messages. It also reduces contention when dispatching to different topics that happen to forward to the same aggregation topic (as happens with all of the channel, bridge and endpoint topics). Since forwards are no longer subscriptions, the bulk of this patch is simply changing stasis_subscription objects to stasis_forward objects (which, admittedly, I should have done in the first place.) Since this required me to yet again put in a growing array, I finally abstracted that out into a set of ast_vector macros in asterisk/vector.h. Review: https://reviewboard.asterisk.org/r/2883/ ........ r400181 | dlee | 2013-09-30 13:48:57 -0500 (Mon, 30 Sep 2013) | 28 lines Remove dispatch object allocation from Stasis publishing While looking for areas for performance improvement, I realized that an unused feature in Stasis was negatively impacting performance. When a message is sent to a subscriber, a dispatch object is allocated for the dispatch, containing the topic the message was published to, the subscriber the message is being sent to, and the message itself. The topic is actually unused by any subscriber in Asterisk today. And the subscriber is associated with the taskprocessor the message is being dispatched to. First, this patch removes the unused topic parameter from Stasis subscription callbacks. Second, this patch introduces the concept of taskprocessor local data, data that may be set on a taskprocessor and provided along with the data pointer when a task is pushed using the ast_taskprocessor_push_local() call. This allows the task to have both data specific to that taskprocessor, in addition to data specific to that invocation. With those two changes, the dispatch object can be removed completely, and the message is simply refcounted and sent directly to the taskprocessor. Review: https://reviewboard.asterisk.org/r/2884/ ........ Merged revisions 399887,400138,400178,400180-400181 from http://svn.asterisk.org/svn/asterisk/branches/12 git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@400186 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2013-09-30 18:55:27 +00:00
struct stasis_message *message)
{
RAII_VAR(struct ast_str *, spyer_channel_string, NULL, ast_free);
RAII_VAR(struct ast_str *, spyee_channel_string, NULL, ast_free);
struct ast_channel_snapshot *spyer;
struct ast_channel_snapshot *spyee;
struct ast_multi_channel_blob *payload = stasis_message_data(message);
spyer = ast_multi_channel_blob_get_channel(payload, "spyer_channel");
if (!spyer) {
ast_log(AST_LOG_WARNING, "Received ChanSpy Start event with no spyer channel!\n");
return;
}
spyee = ast_multi_channel_blob_get_channel(payload, "spyee_channel");
if (!spyee) {
ast_log(AST_LOG_WARNING, "Received ChanSpy Start event with no spyee channel!\n");
return;
}
spyer_channel_string = ast_manager_build_channel_state_string_prefix(spyer, "Spyer");
if (!spyer_channel_string) {
return;
}
spyee_channel_string = ast_manager_build_channel_state_string_prefix(spyee, "Spyee");
if (!spyee_channel_string) {
return;
}
manager_event(EVENT_FLAG_CALL, "ChanSpyStart",
"%s%s",
ast_str_buffer(spyer_channel_string),
ast_str_buffer(spyee_channel_string));
}
static void channel_dtmf_begin_cb(void *data, struct stasis_subscription *sub,
Multiple revisions 399887,400138,400178,400180-400181 ........ r399887 | dlee | 2013-09-26 10:41:47 -0500 (Thu, 26 Sep 2013) | 1 line Minor performance bump by not allocate manager variable struct if we don't need it ........ r400138 | dlee | 2013-09-30 10:24:00 -0500 (Mon, 30 Sep 2013) | 23 lines Stasis performance improvements This patch addresses several performance problems that were found in the initial performance testing of Asterisk 12. The Stasis dispatch object was allocated as an AO2 object, even though it has a very confined lifecycle. This was replaced with a straight ast_malloc(). The Stasis message router was spending an inordinate amount of time searching hash tables. In this case, most of our routers had 6 or fewer routes in them to begin with. This was replaced with an array that's searched linearly for the route. We more heavily rely on AO2 objects in Asterisk 12, and the memset() in ao2_ref() actually became noticeable on the profile. This was #ifdef'ed to only run when AO2_DEBUG was enabled. After being misled by an erroneous comment in taskprocessor.c during profiling, the wrong comment was removed. Review: https://reviewboard.asterisk.org/r/2873/ ........ r400178 | dlee | 2013-09-30 13:26:27 -0500 (Mon, 30 Sep 2013) | 24 lines Taskprocessor optimization; switch Stasis to use taskprocessors This patch optimizes taskprocessor to use a semaphore for signaling, which the OS can do a better job at managing contention and waiting that we can with a mutex and condition. The taskprocessor execution was also slightly optimized to reduce the number of locks taken. The only observable difference in the taskprocessor implementation is that when the final reference to the taskprocessor goes away, it will execute all tasks to completion instead of discarding the unexecuted tasks. For systems where unnamed semaphores are not supported, a really simple semaphore implementation is provided. (Which gives identical performance as the original taskprocessor implementation). The way we ended up implementing Stasis caused the threadpool to be a burden instead of a boost to performance. This was switched to just use taskprocessors directly for subscriptions. Review: https://reviewboard.asterisk.org/r/2881/ ........ r400180 | dlee | 2013-09-30 13:39:34 -0500 (Mon, 30 Sep 2013) | 28 lines Optimize how Stasis forwards are dispatched This patch optimizes how forwards are dispatched in Stasis. Originally, forwards were dispatched as subscriptions that are invoked on the publishing thread. This did not account for the vast number of forwards we would end up having in the system, and the amount of work it would take to walk though the forward subscriptions. This patch modifies Stasis so that rather than walking the tree of forwards on every dispatch, when forwards and subscriptions are changed, the subscriber list for every topic in the tree is changed. This has a couple of benefits. First, this reduces the workload of dispatching messages. It also reduces contention when dispatching to different topics that happen to forward to the same aggregation topic (as happens with all of the channel, bridge and endpoint topics). Since forwards are no longer subscriptions, the bulk of this patch is simply changing stasis_subscription objects to stasis_forward objects (which, admittedly, I should have done in the first place.) Since this required me to yet again put in a growing array, I finally abstracted that out into a set of ast_vector macros in asterisk/vector.h. Review: https://reviewboard.asterisk.org/r/2883/ ........ r400181 | dlee | 2013-09-30 13:48:57 -0500 (Mon, 30 Sep 2013) | 28 lines Remove dispatch object allocation from Stasis publishing While looking for areas for performance improvement, I realized that an unused feature in Stasis was negatively impacting performance. When a message is sent to a subscriber, a dispatch object is allocated for the dispatch, containing the topic the message was published to, the subscriber the message is being sent to, and the message itself. The topic is actually unused by any subscriber in Asterisk today. And the subscriber is associated with the taskprocessor the message is being dispatched to. First, this patch removes the unused topic parameter from Stasis subscription callbacks. Second, this patch introduces the concept of taskprocessor local data, data that may be set on a taskprocessor and provided along with the data pointer when a task is pushed using the ast_taskprocessor_push_local() call. This allows the task to have both data specific to that taskprocessor, in addition to data specific to that invocation. With those two changes, the dispatch object can be removed completely, and the message is simply refcounted and sent directly to the taskprocessor. Review: https://reviewboard.asterisk.org/r/2884/ ........ Merged revisions 399887,400138,400178,400180-400181 from http://svn.asterisk.org/svn/asterisk/branches/12 git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@400186 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2013-09-30 18:55:27 +00:00
struct stasis_message *message)
{
struct ast_channel_blob *obj = stasis_message_data(message);
RAII_VAR(struct ast_str *, channel_event_string, NULL, ast_free);
const char *digit =
ast_json_string_get(ast_json_object_get(obj->blob, "digit"));
const char *direction =
ast_json_string_get(ast_json_object_get(obj->blob, "direction"));
channel_event_string = ast_manager_build_channel_state_string(obj->snapshot);
if (!channel_event_string) {
return;
}
/*** DOCUMENTATION
<managerEvent language="en_US" name="DTMFBegin">
<managerEventInstance class="EVENT_FLAG_DTMF">
<synopsis>Raised when a DTMF digit has started on a channel.</synopsis>
<syntax>
<channel_snapshot/>
<parameter name="Digit">
<para>DTMF digit received or transmitted (0-9, A-E, # or *</para>
</parameter>
<parameter name="Direction">
<enumlist>
<enum name="Received"/>
<enum name="Sent"/>
</enumlist>
</parameter>
</syntax>
<see-also>
<ref type="managerEvent">DTMFEnd</ref>
</see-also>
</managerEventInstance>
</managerEvent>
***/
manager_event(EVENT_FLAG_DTMF, "DTMFBegin",
"%s"
"Digit: %s\r\n"
"Direction: %s\r\n",
ast_str_buffer(channel_event_string),
digit, direction);
}
static void channel_dtmf_end_cb(void *data, struct stasis_subscription *sub,
Multiple revisions 399887,400138,400178,400180-400181 ........ r399887 | dlee | 2013-09-26 10:41:47 -0500 (Thu, 26 Sep 2013) | 1 line Minor performance bump by not allocate manager variable struct if we don't need it ........ r400138 | dlee | 2013-09-30 10:24:00 -0500 (Mon, 30 Sep 2013) | 23 lines Stasis performance improvements This patch addresses several performance problems that were found in the initial performance testing of Asterisk 12. The Stasis dispatch object was allocated as an AO2 object, even though it has a very confined lifecycle. This was replaced with a straight ast_malloc(). The Stasis message router was spending an inordinate amount of time searching hash tables. In this case, most of our routers had 6 or fewer routes in them to begin with. This was replaced with an array that's searched linearly for the route. We more heavily rely on AO2 objects in Asterisk 12, and the memset() in ao2_ref() actually became noticeable on the profile. This was #ifdef'ed to only run when AO2_DEBUG was enabled. After being misled by an erroneous comment in taskprocessor.c during profiling, the wrong comment was removed. Review: https://reviewboard.asterisk.org/r/2873/ ........ r400178 | dlee | 2013-09-30 13:26:27 -0500 (Mon, 30 Sep 2013) | 24 lines Taskprocessor optimization; switch Stasis to use taskprocessors This patch optimizes taskprocessor to use a semaphore for signaling, which the OS can do a better job at managing contention and waiting that we can with a mutex and condition. The taskprocessor execution was also slightly optimized to reduce the number of locks taken. The only observable difference in the taskprocessor implementation is that when the final reference to the taskprocessor goes away, it will execute all tasks to completion instead of discarding the unexecuted tasks. For systems where unnamed semaphores are not supported, a really simple semaphore implementation is provided. (Which gives identical performance as the original taskprocessor implementation). The way we ended up implementing Stasis caused the threadpool to be a burden instead of a boost to performance. This was switched to just use taskprocessors directly for subscriptions. Review: https://reviewboard.asterisk.org/r/2881/ ........ r400180 | dlee | 2013-09-30 13:39:34 -0500 (Mon, 30 Sep 2013) | 28 lines Optimize how Stasis forwards are dispatched This patch optimizes how forwards are dispatched in Stasis. Originally, forwards were dispatched as subscriptions that are invoked on the publishing thread. This did not account for the vast number of forwards we would end up having in the system, and the amount of work it would take to walk though the forward subscriptions. This patch modifies Stasis so that rather than walking the tree of forwards on every dispatch, when forwards and subscriptions are changed, the subscriber list for every topic in the tree is changed. This has a couple of benefits. First, this reduces the workload of dispatching messages. It also reduces contention when dispatching to different topics that happen to forward to the same aggregation topic (as happens with all of the channel, bridge and endpoint topics). Since forwards are no longer subscriptions, the bulk of this patch is simply changing stasis_subscription objects to stasis_forward objects (which, admittedly, I should have done in the first place.) Since this required me to yet again put in a growing array, I finally abstracted that out into a set of ast_vector macros in asterisk/vector.h. Review: https://reviewboard.asterisk.org/r/2883/ ........ r400181 | dlee | 2013-09-30 13:48:57 -0500 (Mon, 30 Sep 2013) | 28 lines Remove dispatch object allocation from Stasis publishing While looking for areas for performance improvement, I realized that an unused feature in Stasis was negatively impacting performance. When a message is sent to a subscriber, a dispatch object is allocated for the dispatch, containing the topic the message was published to, the subscriber the message is being sent to, and the message itself. The topic is actually unused by any subscriber in Asterisk today. And the subscriber is associated with the taskprocessor the message is being dispatched to. First, this patch removes the unused topic parameter from Stasis subscription callbacks. Second, this patch introduces the concept of taskprocessor local data, data that may be set on a taskprocessor and provided along with the data pointer when a task is pushed using the ast_taskprocessor_push_local() call. This allows the task to have both data specific to that taskprocessor, in addition to data specific to that invocation. With those two changes, the dispatch object can be removed completely, and the message is simply refcounted and sent directly to the taskprocessor. Review: https://reviewboard.asterisk.org/r/2884/ ........ Merged revisions 399887,400138,400178,400180-400181 from http://svn.asterisk.org/svn/asterisk/branches/12 git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@400186 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2013-09-30 18:55:27 +00:00
struct stasis_message *message)
{
struct ast_channel_blob *obj = stasis_message_data(message);
RAII_VAR(struct ast_str *, channel_event_string, NULL, ast_free);
const char *digit =
ast_json_string_get(ast_json_object_get(obj->blob, "digit"));
const char *direction =
ast_json_string_get(ast_json_object_get(obj->blob, "direction"));
long duration_ms =
ast_json_integer_get(ast_json_object_get(obj->blob, "duration_ms"));
channel_event_string = ast_manager_build_channel_state_string(obj->snapshot);
if (!channel_event_string) {
return;
}
/*** DOCUMENTATION
<managerEvent language="en_US" name="DTMFEnd">
<managerEventInstance class="EVENT_FLAG_DTMF">
<synopsis>Raised when a DTMF digit has ended on a channel.</synopsis>
<syntax>
<channel_snapshot/>
<parameter name="Digit">
<para>DTMF digit received or transmitted (0-9, A-E, # or *</para>
</parameter>
<parameter name="DurationMs">
<para>Duration (in milliseconds) DTMF was sent/received</para>
</parameter>
<parameter name="Direction">
<enumlist>
<enum name="Received"/>
<enum name="Sent"/>
</enumlist>
</parameter>
</syntax>
<see-also>
<ref type="managerEvent">DTMFBegin</ref>
</see-also>
</managerEventInstance>
</managerEvent>
***/
manager_event(EVENT_FLAG_DTMF, "DTMFEnd",
"%s"
"Digit: %s\r\n"
"DurationMs: %ld\r\n"
"Direction: %s\r\n",
ast_str_buffer(channel_event_string),
digit, duration_ms, direction);
}
static void channel_flash_cb(void *data, struct stasis_subscription *sub,
struct stasis_message *message)
{
struct ast_channel_blob *obj = stasis_message_data(message);
struct ast_str *channel_event_string;
channel_event_string = ast_manager_build_channel_state_string(obj->snapshot);
if (!channel_event_string) {
return;
}
/*** DOCUMENTATION
<managerEvent language="en_US" name="Flash">
<managerEventInstance class="EVENT_FLAG_CALL">
<synopsis>Raised when a hook flash occurs on a channel.</synopsis>
<syntax>
<channel_snapshot/>
</syntax>
</managerEventInstance>
</managerEvent>
***/
manager_event(EVENT_FLAG_CALL, "Flash",
"%s",
ast_str_buffer(channel_event_string));
ast_free(channel_event_string);
}
static void channel_wink_cb(void *data, struct stasis_subscription *sub,
struct stasis_message *message)
{
struct ast_channel_blob *obj = stasis_message_data(message);
struct ast_str *channel_event_string;
channel_event_string = ast_manager_build_channel_state_string(obj->snapshot);
if (!channel_event_string) {
return;
}
/*** DOCUMENTATION
<managerEvent language="en_US" name="Wink">
<managerEventInstance class="EVENT_FLAG_CALL">
<synopsis>Raised when a wink occurs on a channel.</synopsis>
<syntax>
<channel_snapshot/>
</syntax>
</managerEventInstance>
</managerEvent>
***/
manager_event(EVENT_FLAG_CALL, "Wink",
"%s",
ast_str_buffer(channel_event_string));
ast_free(channel_event_string);
}
static void channel_hangup_handler_cb(void *data, struct stasis_subscription *sub,
Multiple revisions 399887,400138,400178,400180-400181 ........ r399887 | dlee | 2013-09-26 10:41:47 -0500 (Thu, 26 Sep 2013) | 1 line Minor performance bump by not allocate manager variable struct if we don't need it ........ r400138 | dlee | 2013-09-30 10:24:00 -0500 (Mon, 30 Sep 2013) | 23 lines Stasis performance improvements This patch addresses several performance problems that were found in the initial performance testing of Asterisk 12. The Stasis dispatch object was allocated as an AO2 object, even though it has a very confined lifecycle. This was replaced with a straight ast_malloc(). The Stasis message router was spending an inordinate amount of time searching hash tables. In this case, most of our routers had 6 or fewer routes in them to begin with. This was replaced with an array that's searched linearly for the route. We more heavily rely on AO2 objects in Asterisk 12, and the memset() in ao2_ref() actually became noticeable on the profile. This was #ifdef'ed to only run when AO2_DEBUG was enabled. After being misled by an erroneous comment in taskprocessor.c during profiling, the wrong comment was removed. Review: https://reviewboard.asterisk.org/r/2873/ ........ r400178 | dlee | 2013-09-30 13:26:27 -0500 (Mon, 30 Sep 2013) | 24 lines Taskprocessor optimization; switch Stasis to use taskprocessors This patch optimizes taskprocessor to use a semaphore for signaling, which the OS can do a better job at managing contention and waiting that we can with a mutex and condition. The taskprocessor execution was also slightly optimized to reduce the number of locks taken. The only observable difference in the taskprocessor implementation is that when the final reference to the taskprocessor goes away, it will execute all tasks to completion instead of discarding the unexecuted tasks. For systems where unnamed semaphores are not supported, a really simple semaphore implementation is provided. (Which gives identical performance as the original taskprocessor implementation). The way we ended up implementing Stasis caused the threadpool to be a burden instead of a boost to performance. This was switched to just use taskprocessors directly for subscriptions. Review: https://reviewboard.asterisk.org/r/2881/ ........ r400180 | dlee | 2013-09-30 13:39:34 -0500 (Mon, 30 Sep 2013) | 28 lines Optimize how Stasis forwards are dispatched This patch optimizes how forwards are dispatched in Stasis. Originally, forwards were dispatched as subscriptions that are invoked on the publishing thread. This did not account for the vast number of forwards we would end up having in the system, and the amount of work it would take to walk though the forward subscriptions. This patch modifies Stasis so that rather than walking the tree of forwards on every dispatch, when forwards and subscriptions are changed, the subscriber list for every topic in the tree is changed. This has a couple of benefits. First, this reduces the workload of dispatching messages. It also reduces contention when dispatching to different topics that happen to forward to the same aggregation topic (as happens with all of the channel, bridge and endpoint topics). Since forwards are no longer subscriptions, the bulk of this patch is simply changing stasis_subscription objects to stasis_forward objects (which, admittedly, I should have done in the first place.) Since this required me to yet again put in a growing array, I finally abstracted that out into a set of ast_vector macros in asterisk/vector.h. Review: https://reviewboard.asterisk.org/r/2883/ ........ r400181 | dlee | 2013-09-30 13:48:57 -0500 (Mon, 30 Sep 2013) | 28 lines Remove dispatch object allocation from Stasis publishing While looking for areas for performance improvement, I realized that an unused feature in Stasis was negatively impacting performance. When a message is sent to a subscriber, a dispatch object is allocated for the dispatch, containing the topic the message was published to, the subscriber the message is being sent to, and the message itself. The topic is actually unused by any subscriber in Asterisk today. And the subscriber is associated with the taskprocessor the message is being dispatched to. First, this patch removes the unused topic parameter from Stasis subscription callbacks. Second, this patch introduces the concept of taskprocessor local data, data that may be set on a taskprocessor and provided along with the data pointer when a task is pushed using the ast_taskprocessor_push_local() call. This allows the task to have both data specific to that taskprocessor, in addition to data specific to that invocation. With those two changes, the dispatch object can be removed completely, and the message is simply refcounted and sent directly to the taskprocessor. Review: https://reviewboard.asterisk.org/r/2884/ ........ Merged revisions 399887,400138,400178,400180-400181 from http://svn.asterisk.org/svn/asterisk/branches/12 git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@400186 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2013-09-30 18:55:27 +00:00
struct stasis_message *message)
{
RAII_VAR(struct ast_str *, channel_event_string, NULL, ast_free);
struct ast_channel_blob *payload = stasis_message_data(message);
const char *action = ast_json_string_get(ast_json_object_get(payload->blob, "type"));
const char *handler = ast_json_string_get(ast_json_object_get(payload->blob, "handler"));
const char *event;
channel_event_string = ast_manager_build_channel_state_string(payload->snapshot);
if (!channel_event_string) {
return;
}
if (!strcmp(action, "run")) {
event = "HangupHandlerRun";
} else if (!strcmp(action, "pop")) {
event = "HangupHandlerPop";
} else if (!strcmp(action, "push")) {
event = "HangupHandlerPush";
} else {
return;
}
manager_event(EVENT_FLAG_DIALPLAN, event,
"%s"
"Handler: %s\r\n",
ast_str_buffer(channel_event_string),
handler);
}
static void channel_fax_cb(void *data, struct stasis_subscription *sub,
Multiple revisions 399887,400138,400178,400180-400181 ........ r399887 | dlee | 2013-09-26 10:41:47 -0500 (Thu, 26 Sep 2013) | 1 line Minor performance bump by not allocate manager variable struct if we don't need it ........ r400138 | dlee | 2013-09-30 10:24:00 -0500 (Mon, 30 Sep 2013) | 23 lines Stasis performance improvements This patch addresses several performance problems that were found in the initial performance testing of Asterisk 12. The Stasis dispatch object was allocated as an AO2 object, even though it has a very confined lifecycle. This was replaced with a straight ast_malloc(). The Stasis message router was spending an inordinate amount of time searching hash tables. In this case, most of our routers had 6 or fewer routes in them to begin with. This was replaced with an array that's searched linearly for the route. We more heavily rely on AO2 objects in Asterisk 12, and the memset() in ao2_ref() actually became noticeable on the profile. This was #ifdef'ed to only run when AO2_DEBUG was enabled. After being misled by an erroneous comment in taskprocessor.c during profiling, the wrong comment was removed. Review: https://reviewboard.asterisk.org/r/2873/ ........ r400178 | dlee | 2013-09-30 13:26:27 -0500 (Mon, 30 Sep 2013) | 24 lines Taskprocessor optimization; switch Stasis to use taskprocessors This patch optimizes taskprocessor to use a semaphore for signaling, which the OS can do a better job at managing contention and waiting that we can with a mutex and condition. The taskprocessor execution was also slightly optimized to reduce the number of locks taken. The only observable difference in the taskprocessor implementation is that when the final reference to the taskprocessor goes away, it will execute all tasks to completion instead of discarding the unexecuted tasks. For systems where unnamed semaphores are not supported, a really simple semaphore implementation is provided. (Which gives identical performance as the original taskprocessor implementation). The way we ended up implementing Stasis caused the threadpool to be a burden instead of a boost to performance. This was switched to just use taskprocessors directly for subscriptions. Review: https://reviewboard.asterisk.org/r/2881/ ........ r400180 | dlee | 2013-09-30 13:39:34 -0500 (Mon, 30 Sep 2013) | 28 lines Optimize how Stasis forwards are dispatched This patch optimizes how forwards are dispatched in Stasis. Originally, forwards were dispatched as subscriptions that are invoked on the publishing thread. This did not account for the vast number of forwards we would end up having in the system, and the amount of work it would take to walk though the forward subscriptions. This patch modifies Stasis so that rather than walking the tree of forwards on every dispatch, when forwards and subscriptions are changed, the subscriber list for every topic in the tree is changed. This has a couple of benefits. First, this reduces the workload of dispatching messages. It also reduces contention when dispatching to different topics that happen to forward to the same aggregation topic (as happens with all of the channel, bridge and endpoint topics). Since forwards are no longer subscriptions, the bulk of this patch is simply changing stasis_subscription objects to stasis_forward objects (which, admittedly, I should have done in the first place.) Since this required me to yet again put in a growing array, I finally abstracted that out into a set of ast_vector macros in asterisk/vector.h. Review: https://reviewboard.asterisk.org/r/2883/ ........ r400181 | dlee | 2013-09-30 13:48:57 -0500 (Mon, 30 Sep 2013) | 28 lines Remove dispatch object allocation from Stasis publishing While looking for areas for performance improvement, I realized that an unused feature in Stasis was negatively impacting performance. When a message is sent to a subscriber, a dispatch object is allocated for the dispatch, containing the topic the message was published to, the subscriber the message is being sent to, and the message itself. The topic is actually unused by any subscriber in Asterisk today. And the subscriber is associated with the taskprocessor the message is being dispatched to. First, this patch removes the unused topic parameter from Stasis subscription callbacks. Second, this patch introduces the concept of taskprocessor local data, data that may be set on a taskprocessor and provided along with the data pointer when a task is pushed using the ast_taskprocessor_push_local() call. This allows the task to have both data specific to that taskprocessor, in addition to data specific to that invocation. With those two changes, the dispatch object can be removed completely, and the message is simply refcounted and sent directly to the taskprocessor. Review: https://reviewboard.asterisk.org/r/2884/ ........ Merged revisions 399887,400138,400178,400180-400181 from http://svn.asterisk.org/svn/asterisk/branches/12 git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@400186 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2013-09-30 18:55:27 +00:00
struct stasis_message *message)
{
RAII_VAR(struct ast_str *, channel_event_string, NULL, ast_free);
RAII_VAR(struct ast_str *, event_buffer, ast_str_create(256), ast_free);
struct ast_channel_blob *payload = stasis_message_data(message);
const char *type = ast_json_string_get(ast_json_object_get(payload->blob, "type"));
struct ast_json *operation = ast_json_object_get(payload->blob, "operation");
struct ast_json *status = ast_json_object_get(payload->blob, "status");
struct ast_json *local_station_id = ast_json_object_get(payload->blob, "local_station_id");
struct ast_json *remote_station_id = ast_json_object_get(payload->blob, "remote_station_id");
struct ast_json *fax_pages = ast_json_object_get(payload->blob, "fax_pages");
struct ast_json *fax_resolution = ast_json_object_get(payload->blob, "fax_resolution");
struct ast_json *fax_bitrate = ast_json_object_get(payload->blob, "fax_bitrate");
struct ast_json *filenames = ast_json_object_get(payload->blob, "filenames");
const char *event;
size_t array_len;
size_t i;
if (!event_buffer) {
return;
}
channel_event_string = ast_manager_build_channel_state_string(payload->snapshot);
if (!channel_event_string) {
return;
}
if (!strcmp(type, "status")) {
event = "FAXStatus";
} else if (!strcmp(type, "receive")) {
event = "ReceiveFAX";
} else if (!strcmp(type, "send")) {
event = "SendFAX";
} else {
return;
}
if (operation) {
ast_str_append(&event_buffer, 0, "Operation: %s\r\n", ast_json_string_get(operation));
}
if (status) {
ast_str_append(&event_buffer, 0, "Status: %s\r\n", ast_json_string_get(status));
}
if (local_station_id) {
ast_str_append(&event_buffer, 0, "LocalStationID: %s\r\n", ast_json_string_get(local_station_id));
}
if (remote_station_id) {
ast_str_append(&event_buffer, 0, "RemoteStationID: %s\r\n", ast_json_string_get(remote_station_id));
}
if (fax_pages) {
ast_str_append(&event_buffer, 0, "PagesTransferred: %s\r\n", ast_json_string_get(fax_pages));
}
if (fax_resolution) {
ast_str_append(&event_buffer, 0, "Resolution: %s\r\n", ast_json_string_get(fax_resolution));
}
if (fax_bitrate) {
ast_str_append(&event_buffer, 0, "TransferRate: %s\r\n", ast_json_string_get(fax_bitrate));
}
if (filenames) {
array_len = ast_json_array_size(filenames);
for (i = 0; i < array_len; i++) {
ast_str_append(&event_buffer, 0, "FileName: %s\r\n", ast_json_string_get(ast_json_array_get(filenames, i)));
}
}
manager_event(EVENT_FLAG_CALL, event,
"%s"
"%s",
ast_str_buffer(channel_event_string),
ast_str_buffer(event_buffer));
}
static void channel_moh_start_cb(void *data, struct stasis_subscription *sub,
Multiple revisions 399887,400138,400178,400180-400181 ........ r399887 | dlee | 2013-09-26 10:41:47 -0500 (Thu, 26 Sep 2013) | 1 line Minor performance bump by not allocate manager variable struct if we don't need it ........ r400138 | dlee | 2013-09-30 10:24:00 -0500 (Mon, 30 Sep 2013) | 23 lines Stasis performance improvements This patch addresses several performance problems that were found in the initial performance testing of Asterisk 12. The Stasis dispatch object was allocated as an AO2 object, even though it has a very confined lifecycle. This was replaced with a straight ast_malloc(). The Stasis message router was spending an inordinate amount of time searching hash tables. In this case, most of our routers had 6 or fewer routes in them to begin with. This was replaced with an array that's searched linearly for the route. We more heavily rely on AO2 objects in Asterisk 12, and the memset() in ao2_ref() actually became noticeable on the profile. This was #ifdef'ed to only run when AO2_DEBUG was enabled. After being misled by an erroneous comment in taskprocessor.c during profiling, the wrong comment was removed. Review: https://reviewboard.asterisk.org/r/2873/ ........ r400178 | dlee | 2013-09-30 13:26:27 -0500 (Mon, 30 Sep 2013) | 24 lines Taskprocessor optimization; switch Stasis to use taskprocessors This patch optimizes taskprocessor to use a semaphore for signaling, which the OS can do a better job at managing contention and waiting that we can with a mutex and condition. The taskprocessor execution was also slightly optimized to reduce the number of locks taken. The only observable difference in the taskprocessor implementation is that when the final reference to the taskprocessor goes away, it will execute all tasks to completion instead of discarding the unexecuted tasks. For systems where unnamed semaphores are not supported, a really simple semaphore implementation is provided. (Which gives identical performance as the original taskprocessor implementation). The way we ended up implementing Stasis caused the threadpool to be a burden instead of a boost to performance. This was switched to just use taskprocessors directly for subscriptions. Review: https://reviewboard.asterisk.org/r/2881/ ........ r400180 | dlee | 2013-09-30 13:39:34 -0500 (Mon, 30 Sep 2013) | 28 lines Optimize how Stasis forwards are dispatched This patch optimizes how forwards are dispatched in Stasis. Originally, forwards were dispatched as subscriptions that are invoked on the publishing thread. This did not account for the vast number of forwards we would end up having in the system, and the amount of work it would take to walk though the forward subscriptions. This patch modifies Stasis so that rather than walking the tree of forwards on every dispatch, when forwards and subscriptions are changed, the subscriber list for every topic in the tree is changed. This has a couple of benefits. First, this reduces the workload of dispatching messages. It also reduces contention when dispatching to different topics that happen to forward to the same aggregation topic (as happens with all of the channel, bridge and endpoint topics). Since forwards are no longer subscriptions, the bulk of this patch is simply changing stasis_subscription objects to stasis_forward objects (which, admittedly, I should have done in the first place.) Since this required me to yet again put in a growing array, I finally abstracted that out into a set of ast_vector macros in asterisk/vector.h. Review: https://reviewboard.asterisk.org/r/2883/ ........ r400181 | dlee | 2013-09-30 13:48:57 -0500 (Mon, 30 Sep 2013) | 28 lines Remove dispatch object allocation from Stasis publishing While looking for areas for performance improvement, I realized that an unused feature in Stasis was negatively impacting performance. When a message is sent to a subscriber, a dispatch object is allocated for the dispatch, containing the topic the message was published to, the subscriber the message is being sent to, and the message itself. The topic is actually unused by any subscriber in Asterisk today. And the subscriber is associated with the taskprocessor the message is being dispatched to. First, this patch removes the unused topic parameter from Stasis subscription callbacks. Second, this patch introduces the concept of taskprocessor local data, data that may be set on a taskprocessor and provided along with the data pointer when a task is pushed using the ast_taskprocessor_push_local() call. This allows the task to have both data specific to that taskprocessor, in addition to data specific to that invocation. With those two changes, the dispatch object can be removed completely, and the message is simply refcounted and sent directly to the taskprocessor. Review: https://reviewboard.asterisk.org/r/2884/ ........ Merged revisions 399887,400138,400178,400180-400181 from http://svn.asterisk.org/svn/asterisk/branches/12 git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@400186 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2013-09-30 18:55:27 +00:00
struct stasis_message *message)
{
struct ast_channel_blob *payload = stasis_message_data(message);
struct ast_json *blob = payload->blob;
RAII_VAR(struct ast_str *, channel_event_string, NULL, ast_free);
channel_event_string = ast_manager_build_channel_state_string(payload->snapshot);
if (!channel_event_string) {
return;
}
manager_event(EVENT_FLAG_CALL, "MusicOnHoldStart",
"%s"
"Class: %s\r\n",
ast_str_buffer(channel_event_string),
ast_json_string_get(ast_json_object_get(blob, "class")));
}
static void channel_moh_stop_cb(void *data, struct stasis_subscription *sub,
Multiple revisions 399887,400138,400178,400180-400181 ........ r399887 | dlee | 2013-09-26 10:41:47 -0500 (Thu, 26 Sep 2013) | 1 line Minor performance bump by not allocate manager variable struct if we don't need it ........ r400138 | dlee | 2013-09-30 10:24:00 -0500 (Mon, 30 Sep 2013) | 23 lines Stasis performance improvements This patch addresses several performance problems that were found in the initial performance testing of Asterisk 12. The Stasis dispatch object was allocated as an AO2 object, even though it has a very confined lifecycle. This was replaced with a straight ast_malloc(). The Stasis message router was spending an inordinate amount of time searching hash tables. In this case, most of our routers had 6 or fewer routes in them to begin with. This was replaced with an array that's searched linearly for the route. We more heavily rely on AO2 objects in Asterisk 12, and the memset() in ao2_ref() actually became noticeable on the profile. This was #ifdef'ed to only run when AO2_DEBUG was enabled. After being misled by an erroneous comment in taskprocessor.c during profiling, the wrong comment was removed. Review: https://reviewboard.asterisk.org/r/2873/ ........ r400178 | dlee | 2013-09-30 13:26:27 -0500 (Mon, 30 Sep 2013) | 24 lines Taskprocessor optimization; switch Stasis to use taskprocessors This patch optimizes taskprocessor to use a semaphore for signaling, which the OS can do a better job at managing contention and waiting that we can with a mutex and condition. The taskprocessor execution was also slightly optimized to reduce the number of locks taken. The only observable difference in the taskprocessor implementation is that when the final reference to the taskprocessor goes away, it will execute all tasks to completion instead of discarding the unexecuted tasks. For systems where unnamed semaphores are not supported, a really simple semaphore implementation is provided. (Which gives identical performance as the original taskprocessor implementation). The way we ended up implementing Stasis caused the threadpool to be a burden instead of a boost to performance. This was switched to just use taskprocessors directly for subscriptions. Review: https://reviewboard.asterisk.org/r/2881/ ........ r400180 | dlee | 2013-09-30 13:39:34 -0500 (Mon, 30 Sep 2013) | 28 lines Optimize how Stasis forwards are dispatched This patch optimizes how forwards are dispatched in Stasis. Originally, forwards were dispatched as subscriptions that are invoked on the publishing thread. This did not account for the vast number of forwards we would end up having in the system, and the amount of work it would take to walk though the forward subscriptions. This patch modifies Stasis so that rather than walking the tree of forwards on every dispatch, when forwards and subscriptions are changed, the subscriber list for every topic in the tree is changed. This has a couple of benefits. First, this reduces the workload of dispatching messages. It also reduces contention when dispatching to different topics that happen to forward to the same aggregation topic (as happens with all of the channel, bridge and endpoint topics). Since forwards are no longer subscriptions, the bulk of this patch is simply changing stasis_subscription objects to stasis_forward objects (which, admittedly, I should have done in the first place.) Since this required me to yet again put in a growing array, I finally abstracted that out into a set of ast_vector macros in asterisk/vector.h. Review: https://reviewboard.asterisk.org/r/2883/ ........ r400181 | dlee | 2013-09-30 13:48:57 -0500 (Mon, 30 Sep 2013) | 28 lines Remove dispatch object allocation from Stasis publishing While looking for areas for performance improvement, I realized that an unused feature in Stasis was negatively impacting performance. When a message is sent to a subscriber, a dispatch object is allocated for the dispatch, containing the topic the message was published to, the subscriber the message is being sent to, and the message itself. The topic is actually unused by any subscriber in Asterisk today. And the subscriber is associated with the taskprocessor the message is being dispatched to. First, this patch removes the unused topic parameter from Stasis subscription callbacks. Second, this patch introduces the concept of taskprocessor local data, data that may be set on a taskprocessor and provided along with the data pointer when a task is pushed using the ast_taskprocessor_push_local() call. This allows the task to have both data specific to that taskprocessor, in addition to data specific to that invocation. With those two changes, the dispatch object can be removed completely, and the message is simply refcounted and sent directly to the taskprocessor. Review: https://reviewboard.asterisk.org/r/2884/ ........ Merged revisions 399887,400138,400178,400180-400181 from http://svn.asterisk.org/svn/asterisk/branches/12 git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@400186 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2013-09-30 18:55:27 +00:00
struct stasis_message *message)
{
struct ast_channel_blob *payload = stasis_message_data(message);
publish_basic_channel_event("MusicOnHoldStop", EVENT_FLAG_CALL, payload->snapshot);
}
static void channel_monitor_start_cb(void *data, struct stasis_subscription *sub,
Multiple revisions 399887,400138,400178,400180-400181 ........ r399887 | dlee | 2013-09-26 10:41:47 -0500 (Thu, 26 Sep 2013) | 1 line Minor performance bump by not allocate manager variable struct if we don't need it ........ r400138 | dlee | 2013-09-30 10:24:00 -0500 (Mon, 30 Sep 2013) | 23 lines Stasis performance improvements This patch addresses several performance problems that were found in the initial performance testing of Asterisk 12. The Stasis dispatch object was allocated as an AO2 object, even though it has a very confined lifecycle. This was replaced with a straight ast_malloc(). The Stasis message router was spending an inordinate amount of time searching hash tables. In this case, most of our routers had 6 or fewer routes in them to begin with. This was replaced with an array that's searched linearly for the route. We more heavily rely on AO2 objects in Asterisk 12, and the memset() in ao2_ref() actually became noticeable on the profile. This was #ifdef'ed to only run when AO2_DEBUG was enabled. After being misled by an erroneous comment in taskprocessor.c during profiling, the wrong comment was removed. Review: https://reviewboard.asterisk.org/r/2873/ ........ r400178 | dlee | 2013-09-30 13:26:27 -0500 (Mon, 30 Sep 2013) | 24 lines Taskprocessor optimization; switch Stasis to use taskprocessors This patch optimizes taskprocessor to use a semaphore for signaling, which the OS can do a better job at managing contention and waiting that we can with a mutex and condition. The taskprocessor execution was also slightly optimized to reduce the number of locks taken. The only observable difference in the taskprocessor implementation is that when the final reference to the taskprocessor goes away, it will execute all tasks to completion instead of discarding the unexecuted tasks. For systems where unnamed semaphores are not supported, a really simple semaphore implementation is provided. (Which gives identical performance as the original taskprocessor implementation). The way we ended up implementing Stasis caused the threadpool to be a burden instead of a boost to performance. This was switched to just use taskprocessors directly for subscriptions. Review: https://reviewboard.asterisk.org/r/2881/ ........ r400180 | dlee | 2013-09-30 13:39:34 -0500 (Mon, 30 Sep 2013) | 28 lines Optimize how Stasis forwards are dispatched This patch optimizes how forwards are dispatched in Stasis. Originally, forwards were dispatched as subscriptions that are invoked on the publishing thread. This did not account for the vast number of forwards we would end up having in the system, and the amount of work it would take to walk though the forward subscriptions. This patch modifies Stasis so that rather than walking the tree of forwards on every dispatch, when forwards and subscriptions are changed, the subscriber list for every topic in the tree is changed. This has a couple of benefits. First, this reduces the workload of dispatching messages. It also reduces contention when dispatching to different topics that happen to forward to the same aggregation topic (as happens with all of the channel, bridge and endpoint topics). Since forwards are no longer subscriptions, the bulk of this patch is simply changing stasis_subscription objects to stasis_forward objects (which, admittedly, I should have done in the first place.) Since this required me to yet again put in a growing array, I finally abstracted that out into a set of ast_vector macros in asterisk/vector.h. Review: https://reviewboard.asterisk.org/r/2883/ ........ r400181 | dlee | 2013-09-30 13:48:57 -0500 (Mon, 30 Sep 2013) | 28 lines Remove dispatch object allocation from Stasis publishing While looking for areas for performance improvement, I realized that an unused feature in Stasis was negatively impacting performance. When a message is sent to a subscriber, a dispatch object is allocated for the dispatch, containing the topic the message was published to, the subscriber the message is being sent to, and the message itself. The topic is actually unused by any subscriber in Asterisk today. And the subscriber is associated with the taskprocessor the message is being dispatched to. First, this patch removes the unused topic parameter from Stasis subscription callbacks. Second, this patch introduces the concept of taskprocessor local data, data that may be set on a taskprocessor and provided along with the data pointer when a task is pushed using the ast_taskprocessor_push_local() call. This allows the task to have both data specific to that taskprocessor, in addition to data specific to that invocation. With those two changes, the dispatch object can be removed completely, and the message is simply refcounted and sent directly to the taskprocessor. Review: https://reviewboard.asterisk.org/r/2884/ ........ Merged revisions 399887,400138,400178,400180-400181 from http://svn.asterisk.org/svn/asterisk/branches/12 git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@400186 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2013-09-30 18:55:27 +00:00
struct stasis_message *message)
{
struct ast_channel_blob *payload = stasis_message_data(message);
publish_basic_channel_event("MonitorStart", EVENT_FLAG_CALL, payload->snapshot);
}
static void channel_monitor_stop_cb(void *data, struct stasis_subscription *sub,
Multiple revisions 399887,400138,400178,400180-400181 ........ r399887 | dlee | 2013-09-26 10:41:47 -0500 (Thu, 26 Sep 2013) | 1 line Minor performance bump by not allocate manager variable struct if we don't need it ........ r400138 | dlee | 2013-09-30 10:24:00 -0500 (Mon, 30 Sep 2013) | 23 lines Stasis performance improvements This patch addresses several performance problems that were found in the initial performance testing of Asterisk 12. The Stasis dispatch object was allocated as an AO2 object, even though it has a very confined lifecycle. This was replaced with a straight ast_malloc(). The Stasis message router was spending an inordinate amount of time searching hash tables. In this case, most of our routers had 6 or fewer routes in them to begin with. This was replaced with an array that's searched linearly for the route. We more heavily rely on AO2 objects in Asterisk 12, and the memset() in ao2_ref() actually became noticeable on the profile. This was #ifdef'ed to only run when AO2_DEBUG was enabled. After being misled by an erroneous comment in taskprocessor.c during profiling, the wrong comment was removed. Review: https://reviewboard.asterisk.org/r/2873/ ........ r400178 | dlee | 2013-09-30 13:26:27 -0500 (Mon, 30 Sep 2013) | 24 lines Taskprocessor optimization; switch Stasis to use taskprocessors This patch optimizes taskprocessor to use a semaphore for signaling, which the OS can do a better job at managing contention and waiting that we can with a mutex and condition. The taskprocessor execution was also slightly optimized to reduce the number of locks taken. The only observable difference in the taskprocessor implementation is that when the final reference to the taskprocessor goes away, it will execute all tasks to completion instead of discarding the unexecuted tasks. For systems where unnamed semaphores are not supported, a really simple semaphore implementation is provided. (Which gives identical performance as the original taskprocessor implementation). The way we ended up implementing Stasis caused the threadpool to be a burden instead of a boost to performance. This was switched to just use taskprocessors directly for subscriptions. Review: https://reviewboard.asterisk.org/r/2881/ ........ r400180 | dlee | 2013-09-30 13:39:34 -0500 (Mon, 30 Sep 2013) | 28 lines Optimize how Stasis forwards are dispatched This patch optimizes how forwards are dispatched in Stasis. Originally, forwards were dispatched as subscriptions that are invoked on the publishing thread. This did not account for the vast number of forwards we would end up having in the system, and the amount of work it would take to walk though the forward subscriptions. This patch modifies Stasis so that rather than walking the tree of forwards on every dispatch, when forwards and subscriptions are changed, the subscriber list for every topic in the tree is changed. This has a couple of benefits. First, this reduces the workload of dispatching messages. It also reduces contention when dispatching to different topics that happen to forward to the same aggregation topic (as happens with all of the channel, bridge and endpoint topics). Since forwards are no longer subscriptions, the bulk of this patch is simply changing stasis_subscription objects to stasis_forward objects (which, admittedly, I should have done in the first place.) Since this required me to yet again put in a growing array, I finally abstracted that out into a set of ast_vector macros in asterisk/vector.h. Review: https://reviewboard.asterisk.org/r/2883/ ........ r400181 | dlee | 2013-09-30 13:48:57 -0500 (Mon, 30 Sep 2013) | 28 lines Remove dispatch object allocation from Stasis publishing While looking for areas for performance improvement, I realized that an unused feature in Stasis was negatively impacting performance. When a message is sent to a subscriber, a dispatch object is allocated for the dispatch, containing the topic the message was published to, the subscriber the message is being sent to, and the message itself. The topic is actually unused by any subscriber in Asterisk today. And the subscriber is associated with the taskprocessor the message is being dispatched to. First, this patch removes the unused topic parameter from Stasis subscription callbacks. Second, this patch introduces the concept of taskprocessor local data, data that may be set on a taskprocessor and provided along with the data pointer when a task is pushed using the ast_taskprocessor_push_local() call. This allows the task to have both data specific to that taskprocessor, in addition to data specific to that invocation. With those two changes, the dispatch object can be removed completely, and the message is simply refcounted and sent directly to the taskprocessor. Review: https://reviewboard.asterisk.org/r/2884/ ........ Merged revisions 399887,400138,400178,400180-400181 from http://svn.asterisk.org/svn/asterisk/branches/12 git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@400186 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2013-09-30 18:55:27 +00:00
struct stasis_message *message)
{
struct ast_channel_blob *payload = stasis_message_data(message);
publish_basic_channel_event("MonitorStop", EVENT_FLAG_CALL, payload->snapshot);
}
static void channel_mixmonitor_start_cb(void *data, struct stasis_subscription *sub,
struct stasis_message *message)
{
struct ast_channel_blob *payload = stasis_message_data(message);
publish_basic_channel_event("MixMonitorStart", EVENT_FLAG_CALL, payload->snapshot);
}
static void channel_mixmonitor_stop_cb(void *data, struct stasis_subscription *sub,
struct stasis_message *message)
{
struct ast_channel_blob *payload = stasis_message_data(message);
publish_basic_channel_event("MixMonitorStop", EVENT_FLAG_CALL, payload->snapshot);
}
static void channel_mixmonitor_mute_cb(void *data, struct stasis_subscription *sub,
struct stasis_message *message)
{
RAII_VAR(struct ast_str *, channel_event_string, NULL, ast_free);
RAII_VAR(struct ast_str *, event_buffer, ast_str_create(64), ast_free);
struct ast_channel_blob *payload = stasis_message_data(message);
struct ast_json *direction = ast_json_object_get(payload->blob, "direction");
const int state = ast_json_is_true(ast_json_object_get(payload->blob, "state"));
if (!event_buffer) {
return;
}
channel_event_string = ast_manager_build_channel_state_string(payload->snapshot);
if (!channel_event_string) {
return;
}
if (direction) {
ast_str_append(&event_buffer, 0, "Direction: %s\r\n", ast_json_string_get(direction));
}
ast_str_append(&event_buffer, 0, "State: %s\r\n", state ? "1" : "0");
manager_event(EVENT_FLAG_CALL, "MixMonitorMute",
"%s"
"%s",
ast_str_buffer(channel_event_string),
ast_str_buffer(event_buffer));
}
static int dial_status_end(const char *dialstatus)
{
return (strcmp(dialstatus, "RINGING") &&
strcmp(dialstatus, "PROCEEDING") &&
strcmp(dialstatus, "PROGRESS"));
}
/*!
* \brief Callback processing messages for channel dialing
*/
static void channel_dial_cb(void *data, struct stasis_subscription *sub,
Multiple revisions 399887,400138,400178,400180-400181 ........ r399887 | dlee | 2013-09-26 10:41:47 -0500 (Thu, 26 Sep 2013) | 1 line Minor performance bump by not allocate manager variable struct if we don't need it ........ r400138 | dlee | 2013-09-30 10:24:00 -0500 (Mon, 30 Sep 2013) | 23 lines Stasis performance improvements This patch addresses several performance problems that were found in the initial performance testing of Asterisk 12. The Stasis dispatch object was allocated as an AO2 object, even though it has a very confined lifecycle. This was replaced with a straight ast_malloc(). The Stasis message router was spending an inordinate amount of time searching hash tables. In this case, most of our routers had 6 or fewer routes in them to begin with. This was replaced with an array that's searched linearly for the route. We more heavily rely on AO2 objects in Asterisk 12, and the memset() in ao2_ref() actually became noticeable on the profile. This was #ifdef'ed to only run when AO2_DEBUG was enabled. After being misled by an erroneous comment in taskprocessor.c during profiling, the wrong comment was removed. Review: https://reviewboard.asterisk.org/r/2873/ ........ r400178 | dlee | 2013-09-30 13:26:27 -0500 (Mon, 30 Sep 2013) | 24 lines Taskprocessor optimization; switch Stasis to use taskprocessors This patch optimizes taskprocessor to use a semaphore for signaling, which the OS can do a better job at managing contention and waiting that we can with a mutex and condition. The taskprocessor execution was also slightly optimized to reduce the number of locks taken. The only observable difference in the taskprocessor implementation is that when the final reference to the taskprocessor goes away, it will execute all tasks to completion instead of discarding the unexecuted tasks. For systems where unnamed semaphores are not supported, a really simple semaphore implementation is provided. (Which gives identical performance as the original taskprocessor implementation). The way we ended up implementing Stasis caused the threadpool to be a burden instead of a boost to performance. This was switched to just use taskprocessors directly for subscriptions. Review: https://reviewboard.asterisk.org/r/2881/ ........ r400180 | dlee | 2013-09-30 13:39:34 -0500 (Mon, 30 Sep 2013) | 28 lines Optimize how Stasis forwards are dispatched This patch optimizes how forwards are dispatched in Stasis. Originally, forwards were dispatched as subscriptions that are invoked on the publishing thread. This did not account for the vast number of forwards we would end up having in the system, and the amount of work it would take to walk though the forward subscriptions. This patch modifies Stasis so that rather than walking the tree of forwards on every dispatch, when forwards and subscriptions are changed, the subscriber list for every topic in the tree is changed. This has a couple of benefits. First, this reduces the workload of dispatching messages. It also reduces contention when dispatching to different topics that happen to forward to the same aggregation topic (as happens with all of the channel, bridge and endpoint topics). Since forwards are no longer subscriptions, the bulk of this patch is simply changing stasis_subscription objects to stasis_forward objects (which, admittedly, I should have done in the first place.) Since this required me to yet again put in a growing array, I finally abstracted that out into a set of ast_vector macros in asterisk/vector.h. Review: https://reviewboard.asterisk.org/r/2883/ ........ r400181 | dlee | 2013-09-30 13:48:57 -0500 (Mon, 30 Sep 2013) | 28 lines Remove dispatch object allocation from Stasis publishing While looking for areas for performance improvement, I realized that an unused feature in Stasis was negatively impacting performance. When a message is sent to a subscriber, a dispatch object is allocated for the dispatch, containing the topic the message was published to, the subscriber the message is being sent to, and the message itself. The topic is actually unused by any subscriber in Asterisk today. And the subscriber is associated with the taskprocessor the message is being dispatched to. First, this patch removes the unused topic parameter from Stasis subscription callbacks. Second, this patch introduces the concept of taskprocessor local data, data that may be set on a taskprocessor and provided along with the data pointer when a task is pushed using the ast_taskprocessor_push_local() call. This allows the task to have both data specific to that taskprocessor, in addition to data specific to that invocation. With those two changes, the dispatch object can be removed completely, and the message is simply refcounted and sent directly to the taskprocessor. Review: https://reviewboard.asterisk.org/r/2884/ ........ Merged revisions 399887,400138,400178,400180-400181 from http://svn.asterisk.org/svn/asterisk/branches/12 git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@400186 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2013-09-30 18:55:27 +00:00
struct stasis_message *message)
{
struct ast_multi_channel_blob *obj = stasis_message_data(message);
const char *dialstatus;
const char *dialstring;
const char *forward;
struct ast_channel_snapshot *caller;
struct ast_channel_snapshot *peer;
RAII_VAR(struct ast_str *, caller_event_string, NULL, ast_free);
RAII_VAR(struct ast_str *, peer_event_string, NULL, ast_free);
caller = ast_multi_channel_blob_get_channel(obj, "caller");
peer = ast_multi_channel_blob_get_channel(obj, "peer");
/* Peer is required - otherwise, who are we dialing? */
ast_assert(peer != NULL);
peer_event_string = ast_manager_build_channel_state_string_prefix(peer, "Dest");
if (!peer_event_string) {
return;
}
if (caller && !(caller_event_string = ast_manager_build_channel_state_string(caller))) {
return;
}
dialstatus = ast_json_string_get(ast_json_object_get(ast_multi_channel_blob_get_json(obj), "dialstatus"));
dialstring = ast_json_string_get(ast_json_object_get(ast_multi_channel_blob_get_json(obj), "dialstring"));
forward = ast_json_string_get(ast_json_object_get(ast_multi_channel_blob_get_json(obj), "forward"));
if (ast_strlen_zero(dialstatus)) {
manager_event(EVENT_FLAG_CALL, "DialBegin",
"%s"
"%s"
"DialString: %s\r\n",
caller_event_string ? ast_str_buffer(caller_event_string) : "",
ast_str_buffer(peer_event_string),
S_OR(dialstring, "unknown"));
} else {
int forwarded = !ast_strlen_zero(forward);
manager_event(EVENT_FLAG_CALL, dial_status_end(dialstatus) ? "DialEnd" : "DialState",
"%s"
"%s"
"%s%s%s"
"DialStatus: %s\r\n",
caller_event_string ? ast_str_buffer(caller_event_string) : "",
ast_str_buffer(peer_event_string),
forwarded ? "Forward: " : "", S_OR(forward, ""), forwarded ? "\r\n" : "",
S_OR(dialstatus, "unknown"));
}
}
static void channel_hold_cb(void *data, struct stasis_subscription *sub,
Multiple revisions 399887,400138,400178,400180-400181 ........ r399887 | dlee | 2013-09-26 10:41:47 -0500 (Thu, 26 Sep 2013) | 1 line Minor performance bump by not allocate manager variable struct if we don't need it ........ r400138 | dlee | 2013-09-30 10:24:00 -0500 (Mon, 30 Sep 2013) | 23 lines Stasis performance improvements This patch addresses several performance problems that were found in the initial performance testing of Asterisk 12. The Stasis dispatch object was allocated as an AO2 object, even though it has a very confined lifecycle. This was replaced with a straight ast_malloc(). The Stasis message router was spending an inordinate amount of time searching hash tables. In this case, most of our routers had 6 or fewer routes in them to begin with. This was replaced with an array that's searched linearly for the route. We more heavily rely on AO2 objects in Asterisk 12, and the memset() in ao2_ref() actually became noticeable on the profile. This was #ifdef'ed to only run when AO2_DEBUG was enabled. After being misled by an erroneous comment in taskprocessor.c during profiling, the wrong comment was removed. Review: https://reviewboard.asterisk.org/r/2873/ ........ r400178 | dlee | 2013-09-30 13:26:27 -0500 (Mon, 30 Sep 2013) | 24 lines Taskprocessor optimization; switch Stasis to use taskprocessors This patch optimizes taskprocessor to use a semaphore for signaling, which the OS can do a better job at managing contention and waiting that we can with a mutex and condition. The taskprocessor execution was also slightly optimized to reduce the number of locks taken. The only observable difference in the taskprocessor implementation is that when the final reference to the taskprocessor goes away, it will execute all tasks to completion instead of discarding the unexecuted tasks. For systems where unnamed semaphores are not supported, a really simple semaphore implementation is provided. (Which gives identical performance as the original taskprocessor implementation). The way we ended up implementing Stasis caused the threadpool to be a burden instead of a boost to performance. This was switched to just use taskprocessors directly for subscriptions. Review: https://reviewboard.asterisk.org/r/2881/ ........ r400180 | dlee | 2013-09-30 13:39:34 -0500 (Mon, 30 Sep 2013) | 28 lines Optimize how Stasis forwards are dispatched This patch optimizes how forwards are dispatched in Stasis. Originally, forwards were dispatched as subscriptions that are invoked on the publishing thread. This did not account for the vast number of forwards we would end up having in the system, and the amount of work it would take to walk though the forward subscriptions. This patch modifies Stasis so that rather than walking the tree of forwards on every dispatch, when forwards and subscriptions are changed, the subscriber list for every topic in the tree is changed. This has a couple of benefits. First, this reduces the workload of dispatching messages. It also reduces contention when dispatching to different topics that happen to forward to the same aggregation topic (as happens with all of the channel, bridge and endpoint topics). Since forwards are no longer subscriptions, the bulk of this patch is simply changing stasis_subscription objects to stasis_forward objects (which, admittedly, I should have done in the first place.) Since this required me to yet again put in a growing array, I finally abstracted that out into a set of ast_vector macros in asterisk/vector.h. Review: https://reviewboard.asterisk.org/r/2883/ ........ r400181 | dlee | 2013-09-30 13:48:57 -0500 (Mon, 30 Sep 2013) | 28 lines Remove dispatch object allocation from Stasis publishing While looking for areas for performance improvement, I realized that an unused feature in Stasis was negatively impacting performance. When a message is sent to a subscriber, a dispatch object is allocated for the dispatch, containing the topic the message was published to, the subscriber the message is being sent to, and the message itself. The topic is actually unused by any subscriber in Asterisk today. And the subscriber is associated with the taskprocessor the message is being dispatched to. First, this patch removes the unused topic parameter from Stasis subscription callbacks. Second, this patch introduces the concept of taskprocessor local data, data that may be set on a taskprocessor and provided along with the data pointer when a task is pushed using the ast_taskprocessor_push_local() call. This allows the task to have both data specific to that taskprocessor, in addition to data specific to that invocation. With those two changes, the dispatch object can be removed completely, and the message is simply refcounted and sent directly to the taskprocessor. Review: https://reviewboard.asterisk.org/r/2884/ ........ Merged revisions 399887,400138,400178,400180-400181 from http://svn.asterisk.org/svn/asterisk/branches/12 git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@400186 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2013-09-30 18:55:27 +00:00
struct stasis_message *message)
{
struct ast_channel_blob *obj = stasis_message_data(message);
struct ast_str *musicclass_string = ast_str_create(32);
struct ast_str *channel_event_string;
if (!musicclass_string) {
return;
}
channel_event_string = ast_manager_build_channel_state_string(obj->snapshot);
if (!channel_event_string) {
ast_free(musicclass_string);
return;
}
if (obj->blob) {
const char *musicclass;
musicclass = ast_json_string_get(ast_json_object_get(obj->blob, "musicclass"));
if (!ast_strlen_zero(musicclass)) {
ast_str_set(&musicclass_string, 0, "MusicClass: %s\r\n", musicclass);
}
}
manager_event(EVENT_FLAG_CALL, "Hold",
"%s"
"%s",
ast_str_buffer(channel_event_string),
ast_str_buffer(musicclass_string));
ast_free(musicclass_string);
ast_free(channel_event_string);
}
static void channel_unhold_cb(void *data, struct stasis_subscription *sub,
Multiple revisions 399887,400138,400178,400180-400181 ........ r399887 | dlee | 2013-09-26 10:41:47 -0500 (Thu, 26 Sep 2013) | 1 line Minor performance bump by not allocate manager variable struct if we don't need it ........ r400138 | dlee | 2013-09-30 10:24:00 -0500 (Mon, 30 Sep 2013) | 23 lines Stasis performance improvements This patch addresses several performance problems that were found in the initial performance testing of Asterisk 12. The Stasis dispatch object was allocated as an AO2 object, even though it has a very confined lifecycle. This was replaced with a straight ast_malloc(). The Stasis message router was spending an inordinate amount of time searching hash tables. In this case, most of our routers had 6 or fewer routes in them to begin with. This was replaced with an array that's searched linearly for the route. We more heavily rely on AO2 objects in Asterisk 12, and the memset() in ao2_ref() actually became noticeable on the profile. This was #ifdef'ed to only run when AO2_DEBUG was enabled. After being misled by an erroneous comment in taskprocessor.c during profiling, the wrong comment was removed. Review: https://reviewboard.asterisk.org/r/2873/ ........ r400178 | dlee | 2013-09-30 13:26:27 -0500 (Mon, 30 Sep 2013) | 24 lines Taskprocessor optimization; switch Stasis to use taskprocessors This patch optimizes taskprocessor to use a semaphore for signaling, which the OS can do a better job at managing contention and waiting that we can with a mutex and condition. The taskprocessor execution was also slightly optimized to reduce the number of locks taken. The only observable difference in the taskprocessor implementation is that when the final reference to the taskprocessor goes away, it will execute all tasks to completion instead of discarding the unexecuted tasks. For systems where unnamed semaphores are not supported, a really simple semaphore implementation is provided. (Which gives identical performance as the original taskprocessor implementation). The way we ended up implementing Stasis caused the threadpool to be a burden instead of a boost to performance. This was switched to just use taskprocessors directly for subscriptions. Review: https://reviewboard.asterisk.org/r/2881/ ........ r400180 | dlee | 2013-09-30 13:39:34 -0500 (Mon, 30 Sep 2013) | 28 lines Optimize how Stasis forwards are dispatched This patch optimizes how forwards are dispatched in Stasis. Originally, forwards were dispatched as subscriptions that are invoked on the publishing thread. This did not account for the vast number of forwards we would end up having in the system, and the amount of work it would take to walk though the forward subscriptions. This patch modifies Stasis so that rather than walking the tree of forwards on every dispatch, when forwards and subscriptions are changed, the subscriber list for every topic in the tree is changed. This has a couple of benefits. First, this reduces the workload of dispatching messages. It also reduces contention when dispatching to different topics that happen to forward to the same aggregation topic (as happens with all of the channel, bridge and endpoint topics). Since forwards are no longer subscriptions, the bulk of this patch is simply changing stasis_subscription objects to stasis_forward objects (which, admittedly, I should have done in the first place.) Since this required me to yet again put in a growing array, I finally abstracted that out into a set of ast_vector macros in asterisk/vector.h. Review: https://reviewboard.asterisk.org/r/2883/ ........ r400181 | dlee | 2013-09-30 13:48:57 -0500 (Mon, 30 Sep 2013) | 28 lines Remove dispatch object allocation from Stasis publishing While looking for areas for performance improvement, I realized that an unused feature in Stasis was negatively impacting performance. When a message is sent to a subscriber, a dispatch object is allocated for the dispatch, containing the topic the message was published to, the subscriber the message is being sent to, and the message itself. The topic is actually unused by any subscriber in Asterisk today. And the subscriber is associated with the taskprocessor the message is being dispatched to. First, this patch removes the unused topic parameter from Stasis subscription callbacks. Second, this patch introduces the concept of taskprocessor local data, data that may be set on a taskprocessor and provided along with the data pointer when a task is pushed using the ast_taskprocessor_push_local() call. This allows the task to have both data specific to that taskprocessor, in addition to data specific to that invocation. With those two changes, the dispatch object can be removed completely, and the message is simply refcounted and sent directly to the taskprocessor. Review: https://reviewboard.asterisk.org/r/2884/ ........ Merged revisions 399887,400138,400178,400180-400181 from http://svn.asterisk.org/svn/asterisk/branches/12 git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@400186 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2013-09-30 18:55:27 +00:00
struct stasis_message *message)
{
struct ast_channel_blob *obj = stasis_message_data(message);
struct ast_str *channel_event_string;
channel_event_string = ast_manager_build_channel_state_string(obj->snapshot);
if (!channel_event_string) {
return;
}
manager_event(EVENT_FLAG_CALL, "Unhold",
"%s",
ast_str_buffer(channel_event_string));
ast_free(channel_event_string);
}
Move more channel events to Stasis; move res_json.c to main/json.c. This patch started out simply as fixing the bouncing tests introduced in r382685, but required some other changes to give it a decent implementation. To fix the bouncing tests, the UserEvent and Newexten AMI events needed to be refactored to dispatch via Stasis. Dispatching directly to AMI resulted in those events sometimes getting ahead of the associated Newchannel events, which would understandably confuse anyone. I found that instead of creating a zillion different message types and structures associated with them, it would be preferable to define a message type that has a channel snapshot and a blob of structured data with a small bit of additional information. The JSON object model provides a very nice way of representing structured data, so I went with that. * Move JSON support from res_json.c to main/json.c * Made libjansson-dev a required dependency * Added an ast_channel_blob message type, which has a channel snapshot and JSON blob of data. * Changed UserEvent and Newexten events so that they are dispatched via ast_channel_blob messages on the channel's topic. * Got rid of the ast_channel_varset message; used ast_channel_blob instead. * Extracted the manager functions converting Stasis channel events to AMI events into manager_channel.c. (issue ASTERISK-21096) Review: https://reviewboard.asterisk.org/r/2381/ git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@383579 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2013-03-22 14:06:46 +00:00
static void manager_channels_shutdown(void)
{
Multiple revisions 399887,400138,400178,400180-400181 ........ r399887 | dlee | 2013-09-26 10:41:47 -0500 (Thu, 26 Sep 2013) | 1 line Minor performance bump by not allocate manager variable struct if we don't need it ........ r400138 | dlee | 2013-09-30 10:24:00 -0500 (Mon, 30 Sep 2013) | 23 lines Stasis performance improvements This patch addresses several performance problems that were found in the initial performance testing of Asterisk 12. The Stasis dispatch object was allocated as an AO2 object, even though it has a very confined lifecycle. This was replaced with a straight ast_malloc(). The Stasis message router was spending an inordinate amount of time searching hash tables. In this case, most of our routers had 6 or fewer routes in them to begin with. This was replaced with an array that's searched linearly for the route. We more heavily rely on AO2 objects in Asterisk 12, and the memset() in ao2_ref() actually became noticeable on the profile. This was #ifdef'ed to only run when AO2_DEBUG was enabled. After being misled by an erroneous comment in taskprocessor.c during profiling, the wrong comment was removed. Review: https://reviewboard.asterisk.org/r/2873/ ........ r400178 | dlee | 2013-09-30 13:26:27 -0500 (Mon, 30 Sep 2013) | 24 lines Taskprocessor optimization; switch Stasis to use taskprocessors This patch optimizes taskprocessor to use a semaphore for signaling, which the OS can do a better job at managing contention and waiting that we can with a mutex and condition. The taskprocessor execution was also slightly optimized to reduce the number of locks taken. The only observable difference in the taskprocessor implementation is that when the final reference to the taskprocessor goes away, it will execute all tasks to completion instead of discarding the unexecuted tasks. For systems where unnamed semaphores are not supported, a really simple semaphore implementation is provided. (Which gives identical performance as the original taskprocessor implementation). The way we ended up implementing Stasis caused the threadpool to be a burden instead of a boost to performance. This was switched to just use taskprocessors directly for subscriptions. Review: https://reviewboard.asterisk.org/r/2881/ ........ r400180 | dlee | 2013-09-30 13:39:34 -0500 (Mon, 30 Sep 2013) | 28 lines Optimize how Stasis forwards are dispatched This patch optimizes how forwards are dispatched in Stasis. Originally, forwards were dispatched as subscriptions that are invoked on the publishing thread. This did not account for the vast number of forwards we would end up having in the system, and the amount of work it would take to walk though the forward subscriptions. This patch modifies Stasis so that rather than walking the tree of forwards on every dispatch, when forwards and subscriptions are changed, the subscriber list for every topic in the tree is changed. This has a couple of benefits. First, this reduces the workload of dispatching messages. It also reduces contention when dispatching to different topics that happen to forward to the same aggregation topic (as happens with all of the channel, bridge and endpoint topics). Since forwards are no longer subscriptions, the bulk of this patch is simply changing stasis_subscription objects to stasis_forward objects (which, admittedly, I should have done in the first place.) Since this required me to yet again put in a growing array, I finally abstracted that out into a set of ast_vector macros in asterisk/vector.h. Review: https://reviewboard.asterisk.org/r/2883/ ........ r400181 | dlee | 2013-09-30 13:48:57 -0500 (Mon, 30 Sep 2013) | 28 lines Remove dispatch object allocation from Stasis publishing While looking for areas for performance improvement, I realized that an unused feature in Stasis was negatively impacting performance. When a message is sent to a subscriber, a dispatch object is allocated for the dispatch, containing the topic the message was published to, the subscriber the message is being sent to, and the message itself. The topic is actually unused by any subscriber in Asterisk today. And the subscriber is associated with the taskprocessor the message is being dispatched to. First, this patch removes the unused topic parameter from Stasis subscription callbacks. Second, this patch introduces the concept of taskprocessor local data, data that may be set on a taskprocessor and provided along with the data pointer when a task is pushed using the ast_taskprocessor_push_local() call. This allows the task to have both data specific to that taskprocessor, in addition to data specific to that invocation. With those two changes, the dispatch object can be removed completely, and the message is simply refcounted and sent directly to the taskprocessor. Review: https://reviewboard.asterisk.org/r/2884/ ........ Merged revisions 399887,400138,400178,400180-400181 from http://svn.asterisk.org/svn/asterisk/branches/12 git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@400186 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2013-09-30 18:55:27 +00:00
stasis_forward_cancel(topic_forwarder);
topic_forwarder = NULL;
Move more channel events to Stasis; move res_json.c to main/json.c. This patch started out simply as fixing the bouncing tests introduced in r382685, but required some other changes to give it a decent implementation. To fix the bouncing tests, the UserEvent and Newexten AMI events needed to be refactored to dispatch via Stasis. Dispatching directly to AMI resulted in those events sometimes getting ahead of the associated Newchannel events, which would understandably confuse anyone. I found that instead of creating a zillion different message types and structures associated with them, it would be preferable to define a message type that has a channel snapshot and a blob of structured data with a small bit of additional information. The JSON object model provides a very nice way of representing structured data, so I went with that. * Move JSON support from res_json.c to main/json.c * Made libjansson-dev a required dependency * Added an ast_channel_blob message type, which has a channel snapshot and JSON blob of data. * Changed UserEvent and Newexten events so that they are dispatched via ast_channel_blob messages on the channel's topic. * Got rid of the ast_channel_varset message; used ast_channel_blob instead. * Extracted the manager functions converting Stasis channel events to AMI events into manager_channel.c. (issue ASTERISK-21096) Review: https://reviewboard.asterisk.org/r/2381/ git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@383579 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2013-03-22 14:06:46 +00:00
}
int manager_channels_init(void)
{
int ret = 0;
struct stasis_topic *manager_topic;
struct stasis_topic *channel_topic;
struct stasis_message_router *message_router;
Move more channel events to Stasis; move res_json.c to main/json.c. This patch started out simply as fixing the bouncing tests introduced in r382685, but required some other changes to give it a decent implementation. To fix the bouncing tests, the UserEvent and Newexten AMI events needed to be refactored to dispatch via Stasis. Dispatching directly to AMI resulted in those events sometimes getting ahead of the associated Newchannel events, which would understandably confuse anyone. I found that instead of creating a zillion different message types and structures associated with them, it would be preferable to define a message type that has a channel snapshot and a blob of structured data with a small bit of additional information. The JSON object model provides a very nice way of representing structured data, so I went with that. * Move JSON support from res_json.c to main/json.c * Made libjansson-dev a required dependency * Added an ast_channel_blob message type, which has a channel snapshot and JSON blob of data. * Changed UserEvent and Newexten events so that they are dispatched via ast_channel_blob messages on the channel's topic. * Got rid of the ast_channel_varset message; used ast_channel_blob instead. * Extracted the manager functions converting Stasis channel events to AMI events into manager_channel.c. (issue ASTERISK-21096) Review: https://reviewboard.asterisk.org/r/2381/ git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@383579 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2013-03-22 14:06:46 +00:00
manager_topic = ast_manager_get_topic();
if (!manager_topic) {
return -1;
}
message_router = ast_manager_get_message_router();
if (!message_router) {
return -1;
}
channel_topic = ast_channel_topic_all();
if (!channel_topic) {
return -1;
Move more channel events to Stasis; move res_json.c to main/json.c. This patch started out simply as fixing the bouncing tests introduced in r382685, but required some other changes to give it a decent implementation. To fix the bouncing tests, the UserEvent and Newexten AMI events needed to be refactored to dispatch via Stasis. Dispatching directly to AMI resulted in those events sometimes getting ahead of the associated Newchannel events, which would understandably confuse anyone. I found that instead of creating a zillion different message types and structures associated with them, it would be preferable to define a message type that has a channel snapshot and a blob of structured data with a small bit of additional information. The JSON object model provides a very nice way of representing structured data, so I went with that. * Move JSON support from res_json.c to main/json.c * Made libjansson-dev a required dependency * Added an ast_channel_blob message type, which has a channel snapshot and JSON blob of data. * Changed UserEvent and Newexten events so that they are dispatched via ast_channel_blob messages on the channel's topic. * Got rid of the ast_channel_varset message; used ast_channel_blob instead. * Extracted the manager functions converting Stasis channel events to AMI events into manager_channel.c. (issue ASTERISK-21096) Review: https://reviewboard.asterisk.org/r/2381/ git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@383579 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2013-03-22 14:06:46 +00:00
}
topic_forwarder = stasis_forward_all(channel_topic, manager_topic);
if (!topic_forwarder) {
Move more channel events to Stasis; move res_json.c to main/json.c. This patch started out simply as fixing the bouncing tests introduced in r382685, but required some other changes to give it a decent implementation. To fix the bouncing tests, the UserEvent and Newexten AMI events needed to be refactored to dispatch via Stasis. Dispatching directly to AMI resulted in those events sometimes getting ahead of the associated Newchannel events, which would understandably confuse anyone. I found that instead of creating a zillion different message types and structures associated with them, it would be preferable to define a message type that has a channel snapshot and a blob of structured data with a small bit of additional information. The JSON object model provides a very nice way of representing structured data, so I went with that. * Move JSON support from res_json.c to main/json.c * Made libjansson-dev a required dependency * Added an ast_channel_blob message type, which has a channel snapshot and JSON blob of data. * Changed UserEvent and Newexten events so that they are dispatched via ast_channel_blob messages on the channel's topic. * Got rid of the ast_channel_varset message; used ast_channel_blob instead. * Extracted the manager functions converting Stasis channel events to AMI events into manager_channel.c. (issue ASTERISK-21096) Review: https://reviewboard.asterisk.org/r/2381/ git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@383579 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2013-03-22 14:06:46 +00:00
return -1;
}
ast_register_cleanup(manager_channels_shutdown);
/* The snapshot type has a special handler as it can result in multiple
* manager events being queued due to aspects of the snapshot itself
* changing.
*/
ret |= stasis_message_router_add(message_router,
ast_channel_snapshot_type(), channel_snapshot_update, NULL);
Move more channel events to Stasis; move res_json.c to main/json.c. This patch started out simply as fixing the bouncing tests introduced in r382685, but required some other changes to give it a decent implementation. To fix the bouncing tests, the UserEvent and Newexten AMI events needed to be refactored to dispatch via Stasis. Dispatching directly to AMI resulted in those events sometimes getting ahead of the associated Newchannel events, which would understandably confuse anyone. I found that instead of creating a zillion different message types and structures associated with them, it would be preferable to define a message type that has a channel snapshot and a blob of structured data with a small bit of additional information. The JSON object model provides a very nice way of representing structured data, so I went with that. * Move JSON support from res_json.c to main/json.c * Made libjansson-dev a required dependency * Added an ast_channel_blob message type, which has a channel snapshot and JSON blob of data. * Changed UserEvent and Newexten events so that they are dispatched via ast_channel_blob messages on the channel's topic. * Got rid of the ast_channel_varset message; used ast_channel_blob instead. * Extracted the manager functions converting Stasis channel events to AMI events into manager_channel.c. (issue ASTERISK-21096) Review: https://reviewboard.asterisk.org/r/2381/ git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@383579 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2013-03-22 14:06:46 +00:00
ret |= stasis_message_router_add(message_router,
ast_channel_dtmf_begin_type(), channel_dtmf_begin_cb, NULL);
ret |= stasis_message_router_add(message_router,
ast_channel_dtmf_end_type(), channel_dtmf_end_cb, NULL);
ret |= stasis_message_router_add(message_router,
ast_channel_flash_type(), channel_flash_cb, NULL);
ret |= stasis_message_router_add(message_router,
ast_channel_wink_type(), channel_wink_cb, NULL);
ret |= stasis_message_router_add(message_router,
ast_channel_hangup_request_type(), channel_hangup_request_cb,
NULL);
Move more channel events to Stasis; move res_json.c to main/json.c. This patch started out simply as fixing the bouncing tests introduced in r382685, but required some other changes to give it a decent implementation. To fix the bouncing tests, the UserEvent and Newexten AMI events needed to be refactored to dispatch via Stasis. Dispatching directly to AMI resulted in those events sometimes getting ahead of the associated Newchannel events, which would understandably confuse anyone. I found that instead of creating a zillion different message types and structures associated with them, it would be preferable to define a message type that has a channel snapshot and a blob of structured data with a small bit of additional information. The JSON object model provides a very nice way of representing structured data, so I went with that. * Move JSON support from res_json.c to main/json.c * Made libjansson-dev a required dependency * Added an ast_channel_blob message type, which has a channel snapshot and JSON blob of data. * Changed UserEvent and Newexten events so that they are dispatched via ast_channel_blob messages on the channel's topic. * Got rid of the ast_channel_varset message; used ast_channel_blob instead. * Extracted the manager functions converting Stasis channel events to AMI events into manager_channel.c. (issue ASTERISK-21096) Review: https://reviewboard.asterisk.org/r/2381/ git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@383579 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2013-03-22 14:06:46 +00:00
ret |= stasis_message_router_add(message_router,
ast_channel_dial_type(), channel_dial_cb, NULL);
ret |= stasis_message_router_add(message_router,
ast_channel_hold_type(), channel_hold_cb, NULL);
ret |= stasis_message_router_add(message_router,
ast_channel_unhold_type(), channel_unhold_cb, NULL);
ret |= stasis_message_router_add(message_router,
ast_channel_fax_type(), channel_fax_cb, NULL);
ret |= stasis_message_router_add(message_router,
ast_channel_chanspy_start_type(), channel_chanspy_start_cb,
NULL);
ret |= stasis_message_router_add(message_router,
ast_channel_chanspy_stop_type(), channel_chanspy_stop_cb, NULL);
ret |= stasis_message_router_add(message_router,
ast_channel_hangup_handler_type(), channel_hangup_handler_cb,
NULL);
ret |= stasis_message_router_add(message_router,
ast_channel_moh_start_type(), channel_moh_start_cb, NULL);
ret |= stasis_message_router_add(message_router,
ast_channel_moh_stop_type(), channel_moh_stop_cb, NULL);
ret |= stasis_message_router_add(message_router,
ast_channel_monitor_start_type(), channel_monitor_start_cb,
NULL);
ret |= stasis_message_router_add(message_router,
ast_channel_monitor_stop_type(), channel_monitor_stop_cb, NULL);
ret |= stasis_message_router_add(message_router,
ast_channel_mixmonitor_start_type(), channel_mixmonitor_start_cb, NULL);
ret |= stasis_message_router_add(message_router,
ast_channel_mixmonitor_stop_type(), channel_mixmonitor_stop_cb, NULL);
ret |= stasis_message_router_add(message_router,
ast_channel_mixmonitor_mute_type(), channel_mixmonitor_mute_cb, NULL);
Move more channel events to Stasis; move res_json.c to main/json.c. This patch started out simply as fixing the bouncing tests introduced in r382685, but required some other changes to give it a decent implementation. To fix the bouncing tests, the UserEvent and Newexten AMI events needed to be refactored to dispatch via Stasis. Dispatching directly to AMI resulted in those events sometimes getting ahead of the associated Newchannel events, which would understandably confuse anyone. I found that instead of creating a zillion different message types and structures associated with them, it would be preferable to define a message type that has a channel snapshot and a blob of structured data with a small bit of additional information. The JSON object model provides a very nice way of representing structured data, so I went with that. * Move JSON support from res_json.c to main/json.c * Made libjansson-dev a required dependency * Added an ast_channel_blob message type, which has a channel snapshot and JSON blob of data. * Changed UserEvent and Newexten events so that they are dispatched via ast_channel_blob messages on the channel's topic. * Got rid of the ast_channel_varset message; used ast_channel_blob instead. * Extracted the manager functions converting Stasis channel events to AMI events into manager_channel.c. (issue ASTERISK-21096) Review: https://reviewboard.asterisk.org/r/2381/ git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@383579 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2013-03-22 14:06:46 +00:00
/* If somehow we failed to add any routes, just shut down the whole
* thing and fail it.
Move more channel events to Stasis; move res_json.c to main/json.c. This patch started out simply as fixing the bouncing tests introduced in r382685, but required some other changes to give it a decent implementation. To fix the bouncing tests, the UserEvent and Newexten AMI events needed to be refactored to dispatch via Stasis. Dispatching directly to AMI resulted in those events sometimes getting ahead of the associated Newchannel events, which would understandably confuse anyone. I found that instead of creating a zillion different message types and structures associated with them, it would be preferable to define a message type that has a channel snapshot and a blob of structured data with a small bit of additional information. The JSON object model provides a very nice way of representing structured data, so I went with that. * Move JSON support from res_json.c to main/json.c * Made libjansson-dev a required dependency * Added an ast_channel_blob message type, which has a channel snapshot and JSON blob of data. * Changed UserEvent and Newexten events so that they are dispatched via ast_channel_blob messages on the channel's topic. * Got rid of the ast_channel_varset message; used ast_channel_blob instead. * Extracted the manager functions converting Stasis channel events to AMI events into manager_channel.c. (issue ASTERISK-21096) Review: https://reviewboard.asterisk.org/r/2381/ git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@383579 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2013-03-22 14:06:46 +00:00
*/
if (ret) {
manager_channels_shutdown();
return -1;
}
return 0;
}