In r388005, macros were introduced to consistently define message
types. This added an assert if a message type was used either before
it was initialized or after it had been cleaned up. It turns out that
this assertion fires during shutdown.
This actually exposed a hidden shutdown ordering problem. Since
unsubscribing is asynchronous, it's possible that the message types
used by the subscription could be freed before the final message of
the subscription was processed.
This patch adds stasis_subscription_join(), which blocks until the
last message has been processed by the subscription. Since joining was
most commonly done right after an unsubscribe, a
stasis_unsubscribe_and_join() convenience function was also added.
Similar functions were also added to the stasis_caching_topic and
stasis_message_router, since they wrap subscriptions and have similar
problems.
Other code in trunk was refactored to join() where appropriate, or at
least verify that the subscription was complete before being
destroyed.
Review: https://reviewboard.asterisk.org/r/2540
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@389011 65c4cc65-6c06-0410-ace0-fbb531ad65f3
Currently, the buffer for processing "inkeys" is limited to 256 characters. If
the user has many keys and the names of those key files are long, the 256
character limit is not enough.
* Change inkeys buffer to be dynamic
(closes issue ASTERISK-21398)
Reported by: Pavel Kopchyk
Tested by: Pavel Kopchyk, Michael L. Young
Patches:
asterisk-21398-iax2-inkeys-dynamic-buffer_v3.diff
by Michael L. Young (license 5026)
Review: https://reviewboard.asterisk.org/r/2501/
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@389009 65c4cc65-6c06-0410-ace0-fbb531ad65f3
This patch does two things:
* It fixes a bug where the outbound channel's application/data set by the
dialing API/app_dial is not communicated until the channel is hung up.
If that happens, AMI would incorrectly send a NewExten event immediately
after a Hangup. This isn't really AMI's fault, as the dialing APIs never
communicated the 'helpful' app/data on the outbound channel until it was
hungup.
* It makes public sending a stasis message about a change in channel state.
This is useful enough that - for now at least - it should be public. If
operations on a channel go to being more coarse-grained, this function
could be made private again.
Review: https://reviewboard.asterisk.org/r/2548
Note that this problem was found and reported by Matt DiMeo.
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@388976 65c4cc65-6c06-0410-ace0-fbb531ad65f3
If DEBUG_THREADS is enabled __ast_rwlock_destroy causes a segfault while trying
to access a possible NULL t->track object. A NULL check has been added before
trying to access the memory.
(closes issue ASTERISK-21724)
Reported by: Corey Farrell
Fixed by: Corey Farrell
Patches:
ast_rwlock_destroy-segv.patch uploaded by Corey Farrell (license 5909)
........
Merged revisions 388838 from http://svn.asterisk.org/svn/asterisk/branches/1.8
........
Merged revisions 388839 from http://svn.asterisk.org/svn/asterisk/branches/11
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@388840 65c4cc65-6c06-0410-ace0-fbb531ad65f3
The snapshot API contains an option that allow for combining of new
and old messages within a single snapshot. New messages, however,
include options beyond just 'INBOX' - it also includes the Urgent
folder. A previous patch that combined INBOX and Urgent accidentally
impacted snapshots that attempted to gain messages from just the Old
folder. This patch fixes the snapshot gathering such that the API
returns the appropriate messages for the folder selected, with and
without the combine option.
This should make it more clear about what's happening.
Review: https://reviewboard.asterisk.org/r/2539/
........
Merged revisions 388816 from http://svn.asterisk.org/svn/asterisk/branches/11
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@388818 65c4cc65-6c06-0410-ace0-fbb531ad65f3
When implementing playback for stasis-http, the monolithicedness of
res_stasis really started to get in my way.
This patch breaks the major components of res_stasis.c into individual
files.
* res/stasis/app.c - Stasis application tracking
* res/stasis/control.c - Channel control objects
* res/stasis/command.c - Channel command object
This refactoring also allows res_stasis applications to be loaded as
independent modules, such as the new res_stasis_answer module.
The bulk of this patch is simply moving code from one file to another,
adjusting names and adding accessors as necessary.
Review: https://reviewboard.asterisk.org/r/2530/
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@388729 65c4cc65-6c06-0410-ace0-fbb531ad65f3
The debug versions of ao2_ref() should only be used if REF_DEBUG is
enabled so nothing is written to /tmp/refs unexpectedly.
(closes issue ASTERISK-21785)
Reported by: abelbeck
Patches:
jira_asterisk_21785_v11.patch (license #5621) patch uploaded by rmudgett
Tested by: abelbeck
........
Merged revisions 388700 from http://svn.asterisk.org/svn/asterisk/branches/11
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@388701 65c4cc65-6c06-0410-ace0-fbb531ad65f3
This moves the JSON event generators out of the Stasis-HTTP modules and
into standalone JSON-related counterparts so that Stasis-HTTP and
res_stasis can depend on them without creating dependency cycles. This
also provides a future location for Swagger Model validator functions
once the generators for that code are written.
Review: https://reviewboard.asterisk.org/r/2534/
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@388668 65c4cc65-6c06-0410-ace0-fbb531ad65f3
The CALL-ID (ie [C-00000074]) is missing when logging to syslog. This was just
an oversight when this feature was added.
* Add CALL-IDs when using syslog
(closes issue ASTERISK-21430)
Reported by: Nikola Ciprich
Tested by: Nikola Ciprich, Michael L. Young
Patches:
asterisk-21430-syslog-callid_trunk.diff by Michael L. Young (license 5026)
Review: https://reviewboard.asterisk.org/r/2526/
........
Merged revisions 388605 from http://svn.asterisk.org/svn/asterisk/branches/11
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@388617 65c4cc65-6c06-0410-ace0-fbb531ad65f3
The prior code committed, r385473, failed to take into consideration that not
all outgoing calls will be to a peer. My fault.
This patch does the following:
* Check if there is a related peer involved. If there is, check and set NAT
settings according to the peer's settings.
* Fix a problem with realtime peers. If the global setting has auto_force_rport
set and we issued a "sip reload" while a peer is still registered, the peer's
flags for NAT are reset to off. When this happens, we were always setting the
contact address of the peer to that of the full contact info that we had.
(closes issue ASTERISK-21374)
Reported by: jmls
Tested by: Michael L. Young
Patches:
asterisk-21374-fix-crash-and-rt-peers.diff by Michael L. Young (license 5026)
Review: https://reviewboard.asterisk.org/r/2524/
........
Merged revisions 388601 from http://svn.asterisk.org/svn/asterisk/branches/11
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@388602 65c4cc65-6c06-0410-ace0-fbb531ad65f3
After the merge of support for the realtime sorcery module, extensions that
contained a pattern were not being found through odbc realtime. It was tracked
down to this one line that was advancing to the next variable list before it
should have been. The removal of this one line fixes this.
Tested this fix on my machine.
Received confirmation that this is the right fix from file on IRC.
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@388375 65c4cc65-6c06-0410-ace0-fbb531ad65f3
I've noticed when doing a graceful shutdown that the res_stasis_http.so
module gets unloaded before the modules that use it, which causes some
asserts during their unload.
While r386928 was a quick hack to get it to not assert and die, this
patch increases the use counts on res_stasis.so and res_stasis_http.so
properly. It's a bigger change than I expected, hence the review instead
of just committing it.
Review: https://reviewboard.asterisk.org/r/2489/
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@388350 65c4cc65-6c06-0410-ace0-fbb531ad65f3
This change adds a framework in res_stasis for handling events from
channel topics. JSON event generation and validation code is created
from event documentation in rest-api/api-docs/events.json to assist in
JSON event generation, ensure consistency, and ensure that accurate
documentation is available for ALL events that are received by
res_stasis applications.
The userevent application has been refactored along with the code that
handles userevent channel blob events to pass the headers as key/value
pairs in the JSON blob. As a side-effect, app_userevent now handles
duplicate keys by overwriting the previous value.
Review: https://reviewboard.asterisk.org/r/2428/
(closes issue ASTERISK-21180)
Patch-By: Kinsey Moore <kmoore@digium.com>
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@388275 65c4cc65-6c06-0410-ace0-fbb531ad65f3
When we send out a CN packet (for instance, in the case of using rtpkeepalives),
we are not setting the payload code properly. Also, we are setting the marker
bit when we shouldn't be according to RFC 3389, section 4.
AST_RTP_CN is not defined by AST_FORMAT codes. Therefore, we should be using
ast_rtp_codecs_payload_code() rather than ast_rtp_codecs_payload_lookup().
11 and trunk already use the appropriate function.
* In 1.8, use ast_rtp_codecs_payload_code()
* Remove the setting of the marker bit
* Fix the debug message by incrementing the seqno after the debug message is set
in order to display the correct seqno that was sent out
(closes issue ASTERISK-21246)
Reported by: Peter Katzmann
Tested by: Peter Katzmann, Michael L. Young
Patches:
asterisk-21246-rtp-cng-payload-error_1.8_v2.diff
uploaded by Michael L. Young (license 5026)
Review: https://reviewboard.asterisk.org/r/2500/
........
Merged revisions 388111 from http://svn.asterisk.org/svn/asterisk/branches/1.8
........
Merged revisions 388112 from http://svn.asterisk.org/svn/asterisk/branches/11
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@388113 65c4cc65-6c06-0410-ace0-fbb531ad65f3
When the "ignorebusy" setting was deprecated, we added some code to allow us to
be compatible with older setups that are still using the "ignorebusy" setting
instead of "ringinuse". We set a char *variable with the column name to use,
which helps the realtime functions to use the correct column in their SQL
queries. When "persistentmembers" is enabled, we are not setting this variable
before the realtime functions were called to load members. This results in the
variable being NULL and therefore causing a segfault when loading members during
the module's process of loading.
The solution was to move the code that sets that variable to be before these
realtime functions are called during the loading of the module.
(closes issue ASTERISK-21738)
Reported by: JoshE
Tested by: JoshE
Patches:
asterisk-21738-rt-ringinuse-field-not-set.diff
uploaded by Michael L. Young (license 5026)
Review: https://reviewboard.asterisk.org/r/2499/
........
Merged revisions 388108 from http://svn.asterisk.org/svn/asterisk/branches/11
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@388110 65c4cc65-6c06-0410-ace0-fbb531ad65f3
A GCC bug[1] can, in some cases, pop up an unsuppressible pedwarn when
using a static inline standard library function from a non-static
inline function.
This normally doesn't show up, but can occur if you're running an
upgrade version of GCC (such as GCC 4.8 on OS X, which normally runs
GCC 4.2).
[1]: http://gcc.gnu.org/bugzilla/show_bug.cgi?id=47816
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@388046 65c4cc65-6c06-0410-ace0-fbb531ad65f3
The <arpa/nameser_compat.h> was introduced way back in OS X Panther, which
itself was end-of-lifed back in 2007. We can assume that any OS X machine
we build on will need that header file :-)
Why bother removing it? The flag we're checking (__APPLE_CC__) is actually
Apple's build number. Self-compiled versions of GCC (such as installing the
latest version of GCC from homebrew) sets the value to 0, making it useless
for this sort of compile flaggery.
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@388045 65c4cc65-6c06-0410-ace0-fbb531ad65f3
Previously, a call to ast_load_realtime_multientry could get away with
passing a NULL parameter to the function, even though it really isn't
supposed to do that. After the change over to using ast_variable instead
of variadic arguments, the realtime engine gets unhappy if you do this.
This was always an unintended function call in app_directory anyway - now,
we just don't call into the realtime function calls if we don't have anything
to query on.
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@388008 65c4cc65-6c06-0410-ace0-fbb531ad65f3
When we first introduced the channel blob types, the JSON blobs were
self identifying by a required "type" field in the JSON object
itself. This, as it turns out, was a bad idea.
When we introduced the message router, it was useless for routing based
on the JSON type. And messages had two type fields to check: the
stasis_message_type() of the message itself, plus the type field in the
JSON blob (but only if it was a blob message).
This patch corrects that mistake by removing the required type field
from JSON blobs, and introducing first class stasis_message_type objects
for the actual message type.
Since we now will have a proliferation of message types, I introduced a
few macros to help reduce the amount of boilerplate necessary to set
them up.
Review: https://reviewboard.asterisk.org/r/2509
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@388005 65c4cc65-6c06-0410-ace0-fbb531ad65f3
An endpoint is an external device/system that may offer/accept
channels to/from Asterisk. While this is a very useful concept for end
users, it is surprisingly not a core concept within Asterisk itself.
This patch defines ast_endpoint as a separate object, which channel
drivers may use to expose their concept of an endpoint. As the channel
driver creates channels, it can use ast_endpoint_add_channel() to
associate channels to the endpoint. This updated the endpoint
appropriately, and forwards all of the channel's events to the
endpoint's topic.
In order to avoid excessive locking on the endpoint object itself, the
mutable state is not accessible via getters. Instead, you can create a
snapshot using ast_endpoint_snapshot_create() to get a consistent
snapshot of the internal state.
This patch also includes a set of topics and messages associated with
endpoints, and implementations of the endpoint-related RESTful
API. chan_sip was updated to create endpoints with SIP peers, but the
state of the endpoints is not updated with the state of the peer.
Along for the ride in this patch is a Stasis test API. This is a
stasis_message_sink object, which can be subscribed to a Stasis
topic. It has functions for blocking while waiting for conditions in
the message sink to be fulfilled.
(closes issue ASTERISK-21421)
Review: https://reviewboard.asterisk.org/r/2492/
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@387932 65c4cc65-6c06-0410-ace0-fbb531ad65f3
RFC6665 4.2.2: ... after a failed State NOTIFY transaction remove the subscription
The problem is that the State Notify requests rely on the 200OK reponse for pacing control
and to not confuse the notify susbsystem.
The issue is, the pendinginvite isn't cleared if a response isn't received,
thus further notify's are never sent.
The solution, follow RFC 6665 4.2.2's 'SHOULD' and remove the subscription after failure.
(closes issue ASTERISK-21677)
Reported by: Dan Martens
Tested by: alecdavis
alecdavis (license 585)
Review https://reviewboard.asterisk.org/r/2475/
........
Merged revisions 387875 from http://svn.asterisk.org/svn/asterisk/branches/1.8
........
Merged revisions 387880 from http://svn.asterisk.org/svn/asterisk/branches/11
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@387885 65c4cc65-6c06-0410-ace0-fbb531ad65f3
Reload support was originally not included for SLA. It was added later,
but in a fairly non-traditional way. It basically sets a flag
indicating that a reload is pending, and then waits for a time where it
thinks everything SLA related is idle and unused, and *then* executes
the reload. It does this because the reload process is destructive. It
starts by throwing everything away and starting over.
There are a number of problems with this approach. One of them is that
the check to see if anything in use was incomplete. This patch makes it
more complete and thus less likely for a crash to occur during reload
processing. However, this approach still has problems so some much more
significant reworking of this code will need to come in as a next step.
Patch credit and testing by CoreDial, LLC.
........
Merged revisions 387688 from http://svn.asterisk.org/svn/asterisk/branches/1.8
........
Merged revisions 387689 from http://svn.asterisk.org/svn/asterisk/branches/11
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@387690 65c4cc65-6c06-0410-ace0-fbb531ad65f3
This change adds the ability for modules to add themselves as observers
to sorcery object types. Observers can be notified when objects are
created, updated, or deleted as well as when the object type is loaded or
reloaded. Observer notifications are done using a thread pool in a serialized
fashion so the caller of the sorcery API calls is minimally impacted.
This also adds the ability to create JSON changesets of a sorcery object.
Tests are also present to confirm all of the above functionality.
Review: https://reviewboard.asterisk.org/r/2477/
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@387662 65c4cc65-6c06-0410-ace0-fbb531ad65f3
This patch:
* Cleans up some doxygen
* Prevents leaking the system level Stasis topics and messages
on exit (users of valgrind will be happier)
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@387633 65c4cc65-6c06-0410-ace0-fbb531ad65f3
