Mark Michelson [Thu, 29 Jan 2015 20:44:07 +0000 (20:44 +0000)]
Use SIPS URIs in Contact headers when appropriate.
RFC 3261 sections 8.1.1.8 and 12.1.1 dictate specific
scenarios when we are required to use SIPS URIs in Contact
headers. Asterisk's non-compliance with this could actually
cause calls to get dropped when communicating with clients
that are strict about checking the Contact header.
Both of the SIP stacks in Asterisk suffered from this issue.
This changeset corrects the behavior in chan_sip.
ASTERISK-24646 #close
Reported by Stephan Eisvogel
Mark Michelson [Thu, 29 Jan 2015 19:52:45 +0000 (19:52 +0000)]
Allow disabling of 100rel support on PJSIP endpoints.
Due to an inversion error, setting 100rel=no would not actually
change the current value of the setting (which defaulted to "yes").
With this fix, the inversion is corrected.
Joshua Colp [Thu, 29 Jan 2015 12:09:23 +0000 (12:09 +0000)]
res_rtp_asterisk: Fix DTLS when used with OpenSSL 1.0.1k
A recent security fix for OpenSSL broke DTLS negotiation for many
applications. This was caused by read ahead not being enabled when it
should be. While a commit has gone into OpenSSL to force read ahead
on for DTLS it may take some time for a release to be made and the
change to be present in distributions (if at all). As enabling read
ahead is a simple one line change this commit does that and fixes
the issue.
ASTERISK-24711 #close
Reported by: Jared Biel
........
Merged revisions 431384 from http://svn.asterisk.org/svn/asterisk/branches/11
Mark Michelson [Wed, 28 Jan 2015 17:37:55 +0000 (17:37 +0000)]
Fix file descriptor leak in RTP code.
SIP requests that offered codecs incompatible with configured values
could result in the allocation of RTP and RTCP ports that would not get
reclaimed later.
ASTERISK-24666 #close
Reported by Y Ateya
Review: https://reviewboard.asterisk.org/r/4323
AST-2015-001
........
Merged revisions 431300 from http://svn.asterisk.org/svn/asterisk/branches/12
Mitigate possible HTTP injection attacks using CURL() function in Asterisk.
CVE-2014-8150 disclosed a vulnerability in libcURL where HTTP request injection
can be performed given properly-crafted URLs.
Since Asterisk makes use of libcURL, and it is possible that users of Asterisk may
get cURL URLs from user input or remote sources, we have made a patch to Asterisk
to prevent such HTTP injection attacks from originating from Asterisk.
Sean Bright [Wed, 28 Jan 2015 12:18:14 +0000 (12:18 +0000)]
media formats: update res_format_attr_opus & silk
In r419044, we changed how formats were handled, but the return value
of the format_parse_sdp_fmtp functions in res_format_attr_opus and
res_format_attr_silk were not updated, causing calls to fail. Ran
into this when getting codec_opus working with Asterisk 13.
Once the return value was corrected, we were crashing in opus_getjoint
because of NULL format attributes. I've fixed this as well in this
patch.
Performing a CLI "module reload" command when there are new pjsip.conf
registration objects defined frequently failed to load them correctly.
What happens is a race condition between res_pjsip pushing its reload into
an asynchronous task processor task and the thread that does the rest of
the reloads when it gets to reloading the res_pjsip_outbound_registration
module. A similar race condition happens between a reload and the CLI/AMI
show registrations commands. The reload updates the current_states
container and the CLI/AMI commands call get_registrations() which builds a
new current_states container.
* Made res_pjsip.c reload_module() use ast_sip_push_task_synchronous()
instead of ast_sip_push_task() to eliminate two threads processing config
reloads at the same time.
* Made get_registrations() not replace the global current_states container
so the CLI/AMI show registrations command cannot interfere with reloading.
You could never add/remove objects in the container without the
possibility of the container being replaced out from under you by
get_registrations().
* Added a registration loaded sorcery instance observer to purge any dead
registration objects since get_registrations() cannot do this job anymore.
The struct ast_sorcery_instance_observer callbacks must be used because
the callback happens inline with the load process. The struct
ast_sorcery_observer callbacks are pushed to a different thread.
* Added some global current_states NULL pointer checks in case the
container disappears because of unload_module().
* Made sorcery's struct ast_sorcery_instance_observer.object_type_loaded
callbacks guaranteed to be called before any struct
ast_sorcery_observer.loaded callbacks will be called.
* Moved the check for non-reloadable objects to before the sorcery
instance loading callbacks happen to short circuit unnecessary work.
Previously with non-reloadable objects, the sorcery instance
loading/loaded callbacks would always happen, the individual wizard
loading/loaded would be prevented, and the non-reloadable type logging
message would be logged for each associated wizard.
Kevin Harwell [Tue, 27 Jan 2015 22:56:39 +0000 (22:56 +0000)]
tcptls: Bad file descriptor error when reloading chan_sip
While running through some scenarios using chan_sip and tcp a problem would
occur that resulted in a flood of bad file descriptor messages on the cli:
tcptls.c:712 ast_tcptls_server_root: Accept failed: Bad file descriptor
The message is received because the underlying socket has been closed, so is
valid. This is probably happening because unloading of chan_sip is not atomic.
That however is outside the scope of this patch. This patch simply stops the
logging of multiple occurrences of that message.
ASTERISK-24728 #close
Reported by: Thomas Thompson
Review: https://reviewboard.asterisk.org/r/4380/
........
Merged revisions 431218 from http://svn.asterisk.org/svn/asterisk/branches/11
Kevin Harwell [Tue, 27 Jan 2015 19:21:08 +0000 (19:21 +0000)]
chan_sip: stale nonce causes failure
When refreshing (with a small expiration) a registration that was sent to
chan_sip the nonce would be considered stale and reject the registration.
What was happening was that the initial registration's "dialog" still existed
in the dialogs container and upon refresh the dialog match algorithm would
choose that as the "dialog" instead of the newly created one. This occurred
because the algorithm did not check to see if the from tag matched if
authentication info was available after the 401. So, it ended up assuming
the original "dialog" was a match and stopped the search. The old "dialog"
of course had an old nonce, thus the stale nonce message.
This fix attempts to leave the original functionality alone except in the case
of a REGISTER. If a REGISTER is received if searches for an existing "dialog"
matching only on the callid. If the expires value is low enough it will reuse
dialog that is there, otherwise it will create a new one.
ASTERISK-24715 #close
Reported by: John Bigelow
Review: https://reviewboard.asterisk.org/r/4367/
........
Merged revisions 431187 from http://svn.asterisk.org/svn/asterisk/branches/11
Kevin Harwell [Tue, 27 Jan 2015 19:08:44 +0000 (19:08 +0000)]
res_pjsip: make it unloadable (take 2)
Due to the original patch causing memory corruptions it was removed until the
problem could be resolved. This patch is the original patch plus some added
locking around stasis router subcription that was needed to avoid the memory
corruption.
Description of the original problem and patch (still applicable):
The res_pjsip module was previously unloadable. With this patch it can now
be unloaded.
This patch is based off the original patch on the issue (listed below) by Corey
Farrell with a few modifications. Namely, removed a few changes not required to
make the module unloadable and also fixed a bug that would cause asterisk to
crash on unloading.
This patch is the first step (should hopefully be followed by another/others at
some point) in allowing res_pjsip and the modules that depend on it to be
unloadable. At this time, res_pjsip and some of the modules that depend on
res_pjsip cannot be unloaded without causing problems of some sort.
The goal of this patch is to get res_pjsip and only res_pjsip to be able to
unload successfully and/or shutdown without incident (crashes, leaks, etc...).
Other dependent modules may still cause problems on unload.
Basically made sure, with the patch applied, that res_pjsip (with no other
dependent modules loaded) could be succesfully unloaded and Asterisk could
shutdown without any leaks or crashes that pertained directly to res_pjsip.
Richard Mudgett [Tue, 27 Jan 2015 17:36:22 +0000 (17:36 +0000)]
app_confbridge: Repeatedly starting and stopping recording ref leaks the recording channel.
Starting and stopping conference recording more than once causes the
recording channels to be leaked. For v13 the channels also show up in the
CLI "core show channels" output.
* Reworked and simplified the recording channel code to use
ast_bridge_impart() instead of managing the recording thread in the
ConfBridge code. The recording channel's ref handling easily falls into
place and other off nominal code paths get handled better as a result.
Joshua Colp [Tue, 27 Jan 2015 17:32:36 +0000 (17:32 +0000)]
bridge / res_pjsip_sdp_rtp: Fix issues with media not being reinvited during direct media.
This change fixes two issues:
1. During a swap operation bridging added the new channel before having the swap channel
leave. This was not handled in bridge_native_rtp and could result in a channel not getting
reinvited back to Asterisk. After this change the swap channel will leave first and the
new channel will then join.
2. If a re-invite was received after a session had been established any upstream elements
(such as bridge_native_rtp) were not notified that they may want to re-evaluate things.
After this change an UPDATE_RTP_PEER control frame is queued when this situation occurs
and upstream can react.
Jonathan Rose [Tue, 27 Jan 2015 17:22:52 +0000 (17:22 +0000)]
Manager: Fix Manager Action ModuleLoad to give correct response when reloading
Prior to this patch, ModuleLoad would respond with an error indicating that
the requested module wasn't found in spite of finding and reloading the
module.
Matthew Jordan [Tue, 27 Jan 2015 17:20:23 +0000 (17:20 +0000)]
ARI: Improve wiki documentation
This patch improves the documentation of ARI on the wiki. Specifically, it
addresses the following:
* Allowed values and allowed ranges weren't documented. This was particularly
frustrating, as Asterisk would reject query parameters with disallowed values
- but we didn't tell anyone what the allowed values were.
* The /play/id operation on /channels and /bridges failed to document all of
the added media resource types.
* Documentation for creating a channel into a Stasis application failed to
note when it occurred, and that creating a channel into Stasis conflicts with
creating a channel into the dialplan.
* Some other minor tweaks in the mustache templates, including italicizing the
parameter type, putting the default value on its own sub-bullet, and some
other nicities.
Matthew Jordan [Tue, 27 Jan 2015 17:10:21 +0000 (17:10 +0000)]
app_confbridge: Restore user's menu name to CLI output of 'confbridge list'
When issuing a 'confbridge list XXXX' CLI command, the resulting output no
longer displays the menu associated with a ConfBridge participant.
The issue was caused by ASTERISK-22760. When that patch was done, it removed
the copying of the menu name associated with the user from the actual user
profile.
This patch fixes the issue by copying the menu name over to the user profile
when the menu hooks are applied to the user. Since that function now does a
little bit more than just apply the hooks, the name of the function has been
changed to cover the copying of the menu name over as well.
In addition, there is a disparity between the menu name length as it is stored
on the conf_menu structure and the confbridge_user structure; this patch makes
the lengths match so that a strcpy can be used.
Joshua Colp [Tue, 27 Jan 2015 11:47:02 +0000 (11:47 +0000)]
res_parking: Fix crash due to race condition when unloading.
There is currently a race condition when unloading the res_parking
module. Depending on the will of the universe the subscription
invocation may occur AFTER the module is unloaded. This is because
the module does NOT use stasis_unsubscribe_and_join when terminating
the subscription. It merely uses stasis_unsubscribe.
This change makes it use stasis_unsubscribe_and_join which is documented
for usage in this exact scenario.
David M. Lee [Mon, 26 Jan 2015 14:49:32 +0000 (14:49 +0000)]
Various fixes for OS X
This patch addresses compilation errors on OS X. It's been a while, so
there's quite a few things.
* Fixed __attribute__ decls in route.h to be portable.
* Fixed htonll and ntohll to work when they are defined as macros.
* Replaced sem_t usage with our ast_sem wrapper.
* Added ast_sem_timedwait to our ast_sem wrapper.
* Fixed some GCC 4.9 warnings using sig*set() functions.
* Fixed some format strings for portability.
* Fixed compilation issues with res_timing_kqueue (although tests still fail
on OS X).
* Fixed menuconfig /sbin/launchd detection, which disables res_timing_kqueue
on OS X).
Matthew Jordan [Sun, 25 Jan 2015 13:42:22 +0000 (13:42 +0000)]
dynamic realtime: Updates fail to work due to update fields being passed over
When a crash was fixed due to usage of the REALTIME function in r423003, a
regression was introduced into ast_update2_realtime where the update fields
passed to the function would be skipped and the lookup field processed twice.
The use of this function is a bit interesting: A variable argument list is
used with two sentinel values - the first marks the end of the lookup
fields/values; the second marks the end of the update fields/values.
Unfortunately, ast_update2_realtime parses over the lookup fields twice, as
opposed to parsing over the update fields. This causes the lookups to succeed,
but the updates itself to have no effect.
Note that the most common instance of this problem occurred in app_voicemail
during the updating of a mailbox password.
Thanks to the issue reporter, Paddy Grice, for pointing out the problem.
Richard Mudgett [Thu, 22 Jan 2015 19:24:28 +0000 (19:24 +0000)]
Bridge core: Pass a ref with the swap channel when joining a bridge.
When code imparts a channel into a bridge to swap with another channel, a
ref needs to be held on the swap channel to ensure that it cannot
dissapear before finding it in the bridge.
* The ast_bridge_join() swap channel parameter now always steals a ref for
the swap channel. This is the only change to the bridge framework's
public API semantics.
* bridge_channel_internal_join() now requires the bridge_channel->swap
channel to pass in a ref.
stasis transfer: fix a race condition on stasis bridge push
After a bridge transfer completes where a local replacement
channel is used, a stasis transfer message with the details
of the transfer is sent. This is processed by stasis which
then sets the stasis app name and replaced channel snapshot
on the replacement channel.
However, since a separate thread was already started to run
stasis on the new replacement channel, a race was on to see
if the message processing would be completed before the app
name was needed, otherwise the channel would be hung up.
This change moves the calls used to set the stasis app name
and the replace snapshot to the bridge_stasis_push function
callback from the bridge transfer logic, allowing the steps
to be completed earlier and more deterministically, and the
race elimianted.
NOTE: the swap channel parameter to bridge_stasis_push (and
thus all bridge push callbacks) must always be present when
performing a swap with another channel.
ASTERISK-24649 #close
Reported by: John Bigelow
Review: https://reviewboard.asterisk.org/r/4341/
Matthew Jordan [Thu, 22 Jan 2015 14:23:13 +0000 (14:23 +0000)]
apps/app_voicemail: Trigger MWI notification with MixMonitor m() option
The MixMonitor m() option allows a recording to be pushed to a specific
voicemail mailbox. If the message is delivered to the mailbox's INBOX, however,
no MWI notification is currently raised.
This patch corrects the issue by properly calling notify_new_state from the
msg_create_from_file function. This will cause MWI to be triggered if the
message was placed in the mailbox's INBOX.
Matthew Jordan [Wed, 21 Jan 2015 13:33:49 +0000 (13:33 +0000)]
channels/chan_sip: Fix registration leak during reload
When the SIP registrations were migrated to using ao2 in what was then trunk,
the explicit destruction of the registrations on module reload was removed and
not replaced with an ao2 equivalent. Debugging done by Stefan Engström, the
issue reporter, on ASTERISK-24673 confirmed that the reference in the
registry_list container was being leaked.
Since the purpose of cleanup_all_regs is to prep a registration for
destruction, this function now calls an ao2_callback function callback with the
OBJ_MULTIPLE | OBJ_NODATA | OBJ_UNLINK flags used to remove the registrations.
This cleans up each registration, and also removes it from the registration
container registry_list.
Review: https://reviewboard.asterisk.org/r/4355/
ASTERISK-24640 #close
Reported by: Max Man
ASTERISK-24673 #close
Reported by: Stefan Engström
Tested by: Stefan Engström
Matthew Jordan [Wed, 21 Jan 2015 13:27:13 +0000 (13:27 +0000)]
AMI: Add documentation for the missing Cdr/CEL events.
This patch adds AMI event documentation for the Cdr and CEL AMI events.
Note that while these events do share fields with each other and with other
channel related events, they do not contain all of the fields in a standard
channel snapshot, nor is the description of the fields identical. As such,
the patch opts for documentation for each field, for each event.
Matthew Jordan [Wed, 21 Jan 2015 13:10:54 +0000 (13:10 +0000)]
apps/app_dial: Don't publish DialEnd twice on unexpected GoSub/Macro values
The Dial application has some interesting options with the mid-call Macro (M)
and GoSub (U) options. If the MACRO_RESULT/GOSUB_RESULT returns specific
values, the Dial application will take some action upon the channels involved
in the dial operation (such as hanging up a particular party, etc.) The Dial
application ensures that a Stasis message is published in the event that
MACRO_RESULT/GOSUB_RESULT returns a value that kills the dial operation, so
that there is a corresponding DialEnd event published in AMI/ARI for the
DialBegin event that preceeded it.
A bug exists where that same DialEnd event will be published on Stasis even if
the value returned in MACRO_RESULT/GOSUB_RESULT is not one that the Dial
application cares about. This causes two DialEnd events to be published - one
with the MACRO_RESULT/GOSUB_RESULT and another with "ANSWERED" - which is all
sorts of wrong.
This patch fixes the bug by ensuring that we only publish a DialEnd message to
Stasis if the Dial application's mid-call Macro/GoSub returns something that
Dial cares about.
Matthew Jordan [Wed, 21 Jan 2015 12:56:49 +0000 (12:56 +0000)]
main/rtp_engine: Format NTP timestamps as unsigned longs
When the RTCP reports are created, the NTP timestamps are stored as strings,
as JSON does not have an integer type long enough to store the value. However,
on 32-bit systems, a signed long may overflow for some portion of the
timestamp.
This patch corrects the overflow by formatting the timestamps as unsigned
longs.
Ashley Sanders [Tue, 20 Jan 2015 16:51:44 +0000 (16:51 +0000)]
ARI: Fixed crash that occurred when updating a bridge when the optional query parameter 'name' was not supplied.
Prior to this changeset, posting to the: /ari/bridges/{bridgeId} endpoint without specifying a value for the [name] query parameter, would crash Asterisk if the bridge you are attempting to create (or update) had the same ID as an existing bridge. The internal mechanism of the POST operation interpreted a null value for name, thus resulting in an error condition that crashed Asterisk.
Richard Mudgett [Tue, 20 Jan 2015 16:46:16 +0000 (16:46 +0000)]
CHANNEL(peer), chan_iax2, res_fax, SNMP agent: Fix deadlock from reaching across a bridge.
Calling ast_channel_bridge_peer() cannot be done while holding any channel
locks. The reported issue hit the deadlock in chan_iax2, but an audit of
the ast_channel_bridge_peer() calls found three more locations where the
same deadlock can occur.
* Made CHANNEL(peer), res_fax, and the SNMP agent not call
ast_channel_bridge_peer() with any channel locked. For CHANNEL(peer) I
had to rework the logic to not hold the channel lock.
* Made chan_iax2 no longer call ast_channel_bridge_peer(). It was done
for legacy reasons that no longer apply.
* Removed the iax.conf forcejitterbuffer option. It is now always enabled
when the jitterbuffer option is enabled. If you put a jitter buffer on a
channel it will be on the channel.
Matthew Jordan [Tue, 20 Jan 2015 02:39:46 +0000 (02:39 +0000)]
contrib/scripts/install_prereq: Don't install 32-bit packages on 64-bit hosts
On Debian based systems, the install_prereq tool uses a search command on
Debian that results in selecting both 64-bit and 32-bit packages. Besides the
waste of disk space, this can actually cause aptitude use 100% of memory on a
VM with 1GB of RAM as it tried to work out all of the 32-bit package
dependencies.
This patch filters out the 32-bit packages on a 64-bit machine, and leaves
32-bit machines alone.
ASTERISK-24048 #close
Reported by: Ben Klang
Tested by: Ben Klang, Matt Jordan
patches:
install_prereq_64-bit_compat.patch uploaded by Ben Klang (License 5876)
........
Merged revisions 430798 from http://svn.asterisk.org/svn/asterisk/branches/11
Matthew Jordan [Tue, 20 Jan 2015 02:32:23 +0000 (02:32 +0000)]
app_voicemail: Temp message left after review/hangup with ODBC/IMAP backend
When using ODBC or IMAP storage, temporary files created on the file system
must be disposed of using the DISPOSE macro. The DELETE macro will map to a
deletion function for the backend storage, but does not clean up any local
files created as a result of the operation.
When using voicemail with the operator and review options enabled, pressing
0 to enter the menu, followed by 1 to save the message, followed by any
other DTMF press to delete the message, will result in the temporary file
lingering on the file system.
This patch properly calls DISPOSE after the DELETE. This causes the local
file to be disposed of.
ASTERISK-24288 #close
Reported by: LEI FU
patches:
voicemail_odbc_review_fix.diff uploaded by LEI FU (License 6640)
........
Merged revisions 430795 from http://svn.asterisk.org/svn/asterisk/branches/11
Mark Michelson [Mon, 19 Jan 2015 18:05:15 +0000 (18:05 +0000)]
Call extension state callbacks at hint creation.
When a hint gets created, any subsequent device or presence
state changes result in extension status events getting sent
out to interested parties. However, at the time of hint creation,
no such event gets sent out, so watchers of extension state are
potentially left in the dark until the first state change after
hint creation.
Joshua Colp [Mon, 19 Jan 2015 13:18:32 +0000 (13:18 +0000)]
res_pjsip / res_pjsip_multihomed: Use the correct transport and addressing information on UAS sessions.
The first thing this patch fixes is UAS dialogs. Previously if a transport was
configured on an endpoint and an inbound session was created there was no guarantee
that requests sent on the dialog would use the correct transport and address
information. This has now been fixed so an explicitly configured transport
is taken into account.
The second thing this patch fixes is res_pjsip_multihomed. The res_pjsip_multihomed
module attempts to determine what transport a message should go out on and what
addressing information should go into the message itself. In a scenario where
multiple transports exist bound to the same IP address but a different port the
code would incorrectly alter the transport and change the message to the wrong
transport. This change makes the res_pjsip_multihomed module smarter so it will
only change the transport and address information in the message when it is
possible and makes sense.
Mark Michelson [Fri, 16 Jan 2015 22:12:25 +0000 (22:12 +0000)]
Fix problem where a hung channel could occur on a failed blind transfer.
Different clients react differently to being told that a blind transfer
has failed. Some will simply send a BYE and be done with it. Others will
attempt to reinvite themselves back onto the call.
In the latter case, we were creating a new channel and then leaving it to
sit forever doing nothing. With this code change, that new channel will
not be created and the dialog with the transferring channel will be cleaned
up properly.
When FAX was developed, apparently the faxregistry.container used to be a
linked list that was converted to an ao2 container. Some of the
replacement ao2 container operations still had explicit lock/unlocks
around them.
Three off nominal code paths in res_fax.c and res_fax_spandsp.c unlock the
channel even though the routine did not lock the channel and other code
paths in the routine do not unlock the channel.
Joshua Colp [Thu, 15 Jan 2015 12:09:40 +0000 (12:09 +0000)]
res_pjsip_outbound_registration: Fix race condition when reloading and listing registrations.
Due to the split of outbound registration state from configuration it is possible during
a reload for a "pjsip show registrations" CLI command to be executed which gets an older
snapshot of the configuration. This configuration may include outbound registrations which
have been removed due to a reload operation occurring at the same time. The code for
printing the outbound registration did not take this into account but now it does.
Matthew Jordan [Thu, 15 Jan 2015 02:18:45 +0000 (02:18 +0000)]
configure: If cross-compiling, assume we have working semaphores
The Asterisk 13 configure.ac checks for HAS_WORKING_SEMAPHORE but does not have
an option for cross-compiling so it fails with an exit. Since we're cross-
compiling, we can't exactly go looking for the header. The semaphore.h header
is relatively common:
* It's part of the POSIX standard
* It's part of GNU C Library
As such, we assume that it will be present when cross-compiling.
As such, this patch defaults "HAS_WORKING_SEMAPHORE" to "1" if cross-compiling
is detected.
If you're cross-compiling to a platform that doesn't support this, then make
sure you re-define this to 0.
Kevin Harwell [Wed, 14 Jan 2015 23:14:47 +0000 (23:14 +0000)]
res_pjsip: make it unloadable
The res_pjsip module was previously unloadable. With this patch it can now
be unloaded.
This patch is based off the original patch on the issue (listed below) by Corey
Farrell with a few modifications. Namely, removed a few changes not required to
make the module unloadable and also fixed a bug that would cause asterisk to
crash on unloading.
This patch is the first step (should hopefully be followed by another/others at
some point) in allowing res_pjsip and the modules that depend on it to be
unloadable. At this time, res_pjsip and some of the modules that depend on
res_pjsip cannot be unloaded without causing problems of some sort.
The goal of this patch is to get res_pjsip and only res_pjsip to be able to
unload successfully and/or shutdown without incident (crashes, leaks, etc...).
Other dependent modules may still cause problems on unload.
Basically made sure, with the patch applied, that res_pjsip (with no other
dependent modules loaded) could be succesfully unloaded and Asterisk could
shutdown without any leaks or crashes that pertained directly to res_pjsip.
Mark Michelson [Wed, 14 Jan 2015 20:27:18 +0000 (20:27 +0000)]
Prevent slow graceful shutdown when outbound publications never started.
The code was missing the case for explicitly destroying an outbound publication
when Asterisk had never actually published anything. The result was that Asterisk
would hang for a while on a graceful shutdown.
With this change, the case is taken into account, and on a graceful shutdown, these
publications are destroyed without the need to actually send a PUBLISH request.
Richard Mudgett [Tue, 13 Jan 2015 18:16:32 +0000 (18:16 +0000)]
app_macro: Don't restore the calling location on a channel redirect.
v11: If a channel redirect to a macro exten of a macro that is active
happens, the redirect location doesn't get executed. Instead the original
macro location is restored and gets reexecuted.
v13: An additional effect happens if a parked call times out to an
extension in the macro that parked the call then the macro is reexecuted
instead of the expected park return location.
* Made not restore the macro calling location on an
AST_SOFTHANGUP_ASYNCGOTO.
* Increased the locked channel range when setting up the macro execution
environment to cover things that should be done while the channel is
locked.
* Removed unnecessary NULL tests before calling ast_free() in
_macro_exec().
Joshua Colp [Tue, 13 Jan 2015 12:06:50 +0000 (12:06 +0000)]
chan_pjsip: Add configure check for 'pjsip_get_dest_info' function.
The 'pjsip_get_dest_info' function is used to determine if the signaling transport
of the dialog is secure or not. This function was added in PJSIP 2.3 and does not
exist in earlier versions.
This configure check allows Asterisk to build and run with older versions at the
loss of the 'secure' argument for the PJSIP CHANNEL dialplan function. Usage of
this argument will require upgrading to PJSIP 2.3.
Richard Mudgett [Mon, 12 Jan 2015 18:34:28 +0000 (18:34 +0000)]
AMI: Revert non-backwards compatible changes from earlier commit.
* Reverted the change to astman_send_listack() to not use the listflag
parameter and always set the value to "Start" so the start capitalization
is consistent. Unfortunately changing the case of a returned value is not
a backward compatible change so for now FAXSessions is going to have to
remain inconsistent with all of the other AMI list actions.
* Reverted the minor protocol error fix in action_getconfig() when no
requested categories are found. Each line needs to be formatted as
"Header: text".
Matthew Jordan [Mon, 12 Jan 2015 18:28:27 +0000 (18:28 +0000)]
configs/samples/features.conf.sample: Document attended transfer DTMF options
The sample config was missing the configuration options for DTMF attended
transfer completion scenarios. The configuration options 'atxferabort',
'atxfercomplete', 'atxferthreeway', and 'atxferswap' are now documented in the
appropriate configuration file.
Matthew Jordan [Mon, 12 Jan 2015 18:01:24 +0000 (18:01 +0000)]
main/syslog: Allow dynamic logs, such as security events, to log to the syslog
The security event log uses a dynamic log level (SECURITY) that is registered
with the Asterisk logging core. Unfortunately, the syslog would ignore log
statements that had a dynamic log level associated with them. Because the
syslog cannot handle ad hoc dynamic log levels, this patch treats any dynamic
log entries sent to the syslog as logs with a level of NOTICE.
ASTERISK-20744 #close
Reported by: Michael Keuter
Tested by: Michael L. Young, Jacek Konieczny
patches:
asterisk-20744-syslog-dynamic-logging_trunk.diff uploaded by Michael L. Young (license 5026)
........
Merged revisions 430506 from http://svn.asterisk.org/svn/asterisk/branches/11
Matthew Jordan [Mon, 12 Jan 2015 15:18:04 +0000 (15:18 +0000)]
funcs/func_curl: Fix memory leak when CURLOPT channel datastore is destroyed
When the channel datastore associated with the usage of CURLOPT on a specific
channel is freed, the underlying structure holding the list of options is not
disposed of. This patch properly frees the structure in the datastore .destroy
callback.
ASTERISK-24672 #close
Reported by: Kristian Hogh
patches:
func_curl-memory-leak.diff uploaded by Kristian Hogh (License 6639)
........
Merged revisions 430487 from http://svn.asterisk.org/svn/asterisk/branches/11
When app_bridge grabs a channel and puts it into
a bridge, the channel should then continue where
it left off in the dialplan after the bridge has
ended. Although it stores the current dialplan
location as an after bridge goto on the channel,
it was executing the same priority again instead
of going to the next priority. By swapping the
"specific" version of bridge_set_after_goto with
bridge_set_after_go_on, the next priority in the
dialplan is executed instead.
ASTERISK-24637 #close
Review: https://reviewboard.asterisk.org/r/4322/
Reported by: John Bigelow
Richard Mudgett [Fri, 9 Jan 2015 17:54:49 +0000 (17:54 +0000)]
AMI: Make AMI actions that generate event lists consistent.
* Made the following AMI actions use list API calls for consistency:
Agents
BridgeInfo
BridgeList
BridgeTechnologyList
ConfbridgeLIst
ConfbridgeLIstRooms
CoreShowChannels
DAHDIShowChannels
DBGet
DeviceStateList
ExtensionStateList
FAXSessions
Hangup
IAXpeerlist
IAXpeers
IAXregistry
MeetmeList
MeetmeListRooms
MWIGet
ParkedCalls
Parkinglots
PJSIPShowEndpoint
PJSIPShowEndpoints
PJSIPShowRegistrationsInbound
PJSIPShowRegistrationsOutbound
PJSIPShowResourceLists
PJSIPShowSubscriptionsInbound
PJSIPShowSubscriptionsOutbound
PresenceStateList
PRIShowSpans
QueueStatus
QueueSummary
ShowDialPlan
SIPpeers
SIPpeerstatus
SIPshowregistry
SKINNYdevices
SKINNYlines
Status
VoicemailUsersList
* Incremented the AMI version to 2.7.0.
* Changed astman_send_listack() to not use the listflag parameter and
always set the value to "Start" so the start capitalization is consistent.
i.e., The FAXSessions used "Start" while the rest of the system used
"start". The corresponding complete event always used "Complete".
* Fixed ami_show_resource_lists() "PJSIPShowResourceLists" to output the
AMI ActionID for all of its list events.
* Fixed off-nominal AMI protocol error in manager_bridge_info(),
manager_parking_status_single_lot(), and
manager_parking_status_all_lots(). Use of astman_send_error() after
responding to the original AMI action request violates the action response
pattern by sending two responses.
* Fixed minor protocol error in action_getconfig() when no requested
categories are found. Each line needs to be formatted as "Header: text".
* Fixed off-nominal memory leak in manager_build_parked_call_string().
* Eliminated unnecessary use of RAII_VAR() in ami_subscription_detail().
Kinsey Moore [Fri, 9 Jan 2015 14:51:39 +0000 (14:51 +0000)]
res_fax: Add T.38 negotiation timeout option
This change makes the T.38 negotiation timeout configurable via
't38timeout' in res_fax.conf or FAXOPT(t38timeout). It was previously
hard coded to be 5000 milliseconds.
This change also handles T.38 switch failures by aborting the fax since
in the case where this can happen, both sides have agreed to switch to
T.38 and Asterisk is unable to do so.
George Joseph [Thu, 8 Jan 2015 21:40:29 +0000 (21:40 +0000)]
res_pjsip_pubsub: Fix persistent subscriptions not surviving graceful shutdown
If you do a 'core (shutdown|restart) graceful' persistent subscriptions won't
survive. If you do a 'core (shutdown|restart) now' or asterisk terminates for
some reason, they do. Here's why...
When asterisk shuts down gracefully, it sends a 'NOTIFY/terminated' to
subscribers for each subscription. This not only tells the subscribers that the
dialog/state machine is done, it also frees the last reference to the
subscription tree which causes the persistent subscription to get deleted from
astdb. When asterisk restarts, nothing's left. Just preventing the delete from
astdb doesn't work because we already told the subscriber to terminate the
dialog so we can't restart it even if it was still in astdb. Everything works
OK if asterisk terminates unexpectedly because we never send the 'terminated'
message so on restart, the subscription is still in astdb and the subscriber is
none the wiser.
This patch suppresses the sending of 'NOTIFY/terminated' on shutdown for
persistent connections.
Tested-by: George Joseph
Review: https://reviewboard.asterisk.org/r/4318/
Every time a registration started, sip_outbound_registration_response_cb bumps
the ref count on client_state then pushes a handle_registration_response task.
handle_registration_response never unreffed it though. So every time a
registration goes out, the ref count goes up by one.
This patch adds the unreffs to handle_registration_response.
Tested-by: George Joseph
Review: https://reviewboard.asterisk.org/r/4303/
George Joseph [Thu, 8 Jan 2015 17:48:29 +0000 (17:48 +0000)]
res_pjsip_outbound_registration: Fix several reload issues
There are 2 issues with reloading registrations...
1. The 'can_reuse_registration' test wasn't considering the intervals or
expiration in its determination of whether a registration changed or not so if
you changed any of the intervals or the expiration and reloaded, the object
would get reloaded but the actual timers wouldn't change.
can_reuse_registration now does a sorcery diff on the old and new objects
instead of discretely testing certain fields. Now if you change expiration for
instance, and reload, the timer is updated and re-registration will occur on the
new value.
2. If you mung up your password on an outbound registration you get a permanent
failure. If you fix the password (on the outbound_auth object) and reload,
nothing tells outbound_registration to try again because the registration itself
didn't change. This patch adds an observer on the "auth" object type and if any
auth changes, existing registration states are searched and those in a
REJECTED_PERMANENT state are retried.
Tested-by: George Joseph
Review: https://reviewboard.asterisk.org/r/4304/
Kinsey Moore [Wed, 7 Jan 2015 21:25:34 +0000 (21:25 +0000)]
ARI: Allow usage of ASYNCGOTO with Stasis()
When the AMI Redirect action is used with a channel bridged inside
Stasis() and not running a pbx, the channel is hung up instead of
proceeding to the desired location in dialplan. This change allows
such channels to be Redirected properly by detecting the operation
used by Redirect (ASYNCGOTO) and using the code already established
for functionality of the ARI channel continue operation.
George Joseph [Wed, 7 Jan 2015 18:17:05 +0000 (18:17 +0000)]
res_pjsip_exten_state: Change 'does not exist' warning to notice
The 'new_subscribe: Extension <> does not exist or has no associated hint'
is a config issue and doesn't need to clutter up logs with warnings.
Changed to notice.
Tested-by: George Joseph
Review: https://reviewboard.asterisk.org/r/4307/
George Joseph [Wed, 7 Jan 2015 18:14:23 +0000 (18:14 +0000)]
res_pjsip_mwi: Change "MWI Subscription failed" message from warning to notice
The "MWI Subscription failed" message means the client is trying to subscribe
to a mailbox that doesn't exist. There's no need to clutter up logs with
warnings for a client misconfiguration so I changed it to a notice.
Tested-by: George Joseph
Review: https://reviewboard.asterisk.org/r/4306/
George Joseph [Wed, 7 Jan 2015 17:51:35 +0000 (17:51 +0000)]
func_config: Add ability to retrieve specific occurrence of a variable
I guess nobody uses templates with AST_CONFIG because today if you have a
context that inherits from a template and you call AST_CONFIG on the context,
you'll get the value from the template even if you've overridden it in the
context. This is because AST_CONFIG only gets the first occurrence which is
always from the template.
This patch adds an optional 'index' parameter to AST_CONFIG which lets you
specify the exact occurrence to retrieve, or '-1' to retrieve the last.
The default behavior is the current behavior.
Tested-by: George Joseph
Review: https://reviewboard.asterisk.org/r/4313/
Mark Michelson [Wed, 7 Jan 2015 17:35:27 +0000 (17:35 +0000)]
Fix ability to perform a remote attended transfer with PJSIP.
This fix has two parts:
* Corrected an error message to properly state that external_replaces is an extension. The
error message also prints what dialplan context the external_replaces extension was being
looked for in.
* Corrected the printing of the Replaces: header in an INVITE request. We were duplicating
"Replaces: " in the header.
George Joseph [Wed, 7 Jan 2015 16:55:14 +0000 (16:55 +0000)]
config: Add option to NOT preserve effective context when changing a template
Let's say you have a template T with variable VAR1 = ON and you have a
context C(T) that doesn't specify VAR1. If you read C, the effective value
of VAR1 is ON. Now you change T VAR1 to OFF and call
ast_config_text_file_save. The current behavior is that the file gets
re-written with T/VAR1=OFF but C/VAR1=ON is added. Personally, I think this
is a bug. It's preserving the effective state of C even though I didn't
specify C/VAR1 in th first place. I believe the behavior should be that if
I didn't specify C/VAR1 originally, then the effective value of C/VAR1 should
continue to follow the inherited state. Now, if I DID explicitly specify
C/VAR1, the it should be preserved even if the template changes.
Even though I think the existing behavior is a bug, it's been that way forever
so I'm not changing it. Instead, I've created ast_config_text_file_save2()
that takes a bitmask of flags, one of which is to preserve the effective context
(the current behavior). The original ast_config_text_file_save calls *2 with
the preserve flag. If you want the new behavior, call *2 directly without a
flag.
I've also updated Manager UpdateConfig with a new parameter
'PreserveEffectiveContext' whose default is 'yes'. If you want the new behavior
with UpdateConfig, set 'PreserveEffectiveContext: no'.
Tested-by: George Joseph
Review: https://reviewboard.asterisk.org/r/4297/
Matthew Jordan [Tue, 6 Jan 2015 22:46:07 +0000 (22:46 +0000)]
contrib/ast-db-manage: Correct down_revision path for user_eq_phone
When the user_eq_phone patch was backported to 13, it referenced the downward
revision that the PJSIP optimistic encryption option also references. This
creates a multi-path upgrade Exception when generating the SQL files.
This patch corrects this in the 13 branch. Note that trunk, which already
contained both of these features, is unaffected by this problem.
George Joseph [Tue, 6 Jan 2015 17:52:41 +0000 (17:52 +0000)]
res_pjsip_mwi: Change warning to notice
When res_pjsip loads and an endpoint auto-subscribes a mailbox for mwi,
if a contact hasn't registered yet, res_pjsip_mwi spits out a warning.
This is a perfectly normal situation though and doesn't require something
as serious as a warning. It's also self correcting. The device will start
getting mwi as soon as it registers.
This patch changes the warning to a notice.
Tested-by: George Joseph
Review: https://reviewboard.asterisk.org/r/4314/
George Joseph [Tue, 6 Jan 2015 17:35:21 +0000 (17:35 +0000)]
outbound_registration: Add 'pjsip send register' and update 'send unregister'
The current behavior of 'pjsip send unregister' is to send the unregister
(REGISTER with 0 exp) but let the next scheduled register proceed normally.
I don't think that's a good idea. If you unregister, it should stay
unregistered until you decide to start registrations again. So this patch
just adds a cancel_registration call to the current unregister_task to
cancel the timer.
Of course, now you need a way to start registration again so I've added
a 'pjsip send register' command that unregisters and cancels any existing
registration (the same as send unregister), then sends an immediate
registration and starts the timer back up again.
Both changes also ripple to AMI. There's a new PJSIPRegister command.
There's no harm in calling either command repeatedly. They don't care
about the actual state.
Tested-by: George Joseph
Review: https://reviewboard.asterisk.org/r/4301/
George Joseph [Tue, 6 Jan 2015 17:28:40 +0000 (17:28 +0000)]
pjsip cli: Fix sorting of contacts for 'pjsip list contacts'
For some reason I was using a hash container instead of a list to gather the
contacts for 'pjsip list/show contacts' so even though I had a sort function,
the output wasn't sorted. This patch just changes the hash container to a
list container and the contacts now appear sorted in the CLI.
Tested-by: George Joseph
Review: https://reviewboard.asterisk.org/r/4305/
bridge: avoid leaking channel during blond transfer pt2
A blond transfer to a failed destination, when followed
by a recall attempt, lead to a leak of the reference to
the destination channel. In addition to correcting the
regression on the previous attempt (r429826) this fixes
the leak and two additional reference leaks on failures
of bridge_import.
Joshua Colp [Mon, 5 Jan 2015 17:51:59 +0000 (17:51 +0000)]
pjsip: Add 'PJSIP_AOR' and 'PJSIP_CONTACT' dialplan functions.
The PJSIP_AOR dialplan function allows inspection of configured AORs including
what contacts are currently bound to them.
The PJSIP_CONTACT dialplan function allows inspection of contacts in existence.
These can include both externally added (by way of registration) or permanent
ones.
Kinsey Moore [Mon, 29 Dec 2014 13:10:56 +0000 (13:10 +0000)]
PJSIP: Update transport method documentation
This updates the documentation for the 'method' configuration option to
be more verbose about the behaviors of values 'unspecified' and
'default'. They do exactly the same thing which is to select the
default as defined by PJSIP which is currently TLSv1.
Matthew Jordan [Wed, 24 Dec 2014 13:25:04 +0000 (13:25 +0000)]
res_pjsip_keepalive: Add runtime configurable keepalive module for connection-oriented transports.
Note that this is backport from trunk of r425825.
This change adds a module which is configurable using the keep_alive_interval setting in the
global section that will send a CRLF keep alive to all active connection-oriented transports at
the provided interval. This is useful because it can help keep connections open through NATs.
This functionality also exists within PJSIP but can not be controlled at runtime and requires
recompiling it.
Matthew Jordan [Wed, 24 Dec 2014 13:20:00 +0000 (13:20 +0000)]
res_pjsip: Add 'user_eq_phone' option to add a 'user=phone' parameter when applicable.
Note that this is a backport of r425804 from trunk.
This change adds a configuration option which adds a 'user=phone' parameter if the user
portion of the request URI or the From URI is determined to be a number.
George Joseph [Tue, 23 Dec 2014 23:18:50 +0000 (23:18 +0000)]
pjsip_options: Fix continued qualifies after endpoint/aor deletion
If you remove an endpoint/aor from pjsip.conf then do a core reload,
qualifies will continue even though the object are gone. This happens
because nothing clears out the qualify tasks.
This patch unschedules all existing qualify tasks before scheduling
new ones on reload.
Tested-by: George Joseph
Review: https://reviewboard.asterisk.org/r/4290/
Richard Mudgett [Mon, 22 Dec 2014 21:18:22 +0000 (21:18 +0000)]
DTMF atxfer: Setup recall channels as if the transferee initiated the call.
After the initial DTMF atxfer call attempt to the transfer target fails to
answer during a blonde transfer, the recall callback channels do not get
setup with information from the initial transferrer channel. As a result,
the recall callback to the transferrer does not have callid, channel
variables, datastores, accountcode, peeraccount, COLP, and CLID setup. A
similar situation happens with the recall callback to the transfer target
but it is less visible. The recall callback to the transfer target does
not have callid, channel variables, datastores, accountcode, peeraccount,
and COLP setup.
* Added missing information to the recall callback channels before
initiating the call. callid, channel variables, datastores, accountcode,
peeraccount, COLP, and CLID
* Set callid of the transferrer channel on the DTMF atxfer controller
thread attended_transfer_monitor_thread().
* Added missing channel unlocks and props unref to off nominal paths in
attended_transfer_properties_alloc().
ASTERISK-23841 #close
Reported by: Richard Mudgett
Richard Mudgett [Mon, 22 Dec 2014 19:44:03 +0000 (19:44 +0000)]
queue_log: Post QUEUESTART entry when Asterisk fully boots.
The QUEUESTART log entry has historically acted like a fully booted event
for the queue_log file. When the QUEUESTART entry was posted to the log
was broken by the change made by ASTERISK-15863.
* Made post the QUEUESTART queue_log entry when Asterisk fully boots.
This restores the intent of that log entry and happens after realtime has
had a chance to load.
Matthew Jordan [Mon, 22 Dec 2014 15:40:06 +0000 (15:40 +0000)]
chan_sip: Send CANCEL via original INVITE destination even after UPDATE request
Given the following scenario:
* Three SIP phones (A, B, C), all communicating via a proxy with Asterisk
* A call is established between A and B. B performs a SIP attended transfer of
A to C. B sets the call on hold (A is hearing MOH) and dials the extension of
C. While phone C is ringing, B transfers the call (that is, what we typically
call a 'blond transfer').
* When the transfer completes, A hears the ringing of phone C, while B is idle.
In the SIP messaging for the above scenario, a REFER request is sent to
transfer the call. When "sendrpid=yes" is set in sip.conf, Asterisk may send an
UPDATE request to phone C to update party information. This update is sent
directly to phone C, not through the intervening proxy. This has the unfortunate
side effect of providing route information, which is then set on the sip_pvt
structure for C. If someone (e.g. B) is trying to get the call back (through a
directed pickup), Asterisk will send a CANCEL request to C. However, since we
have now updated the route set, the CANCEL request will be sent directly to C
and not through the proxy. The phone ignores this CANCEL according to RFC3261
(Section 9.1).
This patch updates reqprep such that the route is not updated if an UPDATE
request is being sent while the INVITE state is INV_PROCEEDING or
INV_EARLY_MEDIA. This ensures that a subsequent CANCEL request is still sent
to the correct location.
George Joseph [Mon, 22 Dec 2014 00:17:02 +0000 (00:17 +0000)]
res_pjsip_phoneprovi_provider: Fix reload
Reloading wasn't working correctly because on a reload, the sorcery apply
handler was never being called for unchanged users. So, instead of using
an apply handler, I'm now iterating over all users. Works much more reliably.
Tested-by: George Joseph
Review: https://reviewboard.asterisk.org/r/4288/
Joshua Colp [Sat, 20 Dec 2014 20:57:19 +0000 (20:57 +0000)]
acl: Fix reloading of configuration if configuration file does not exist at startup.
The named ACL code incorrectly destroyed the config options information if loading
of the configuration file failed at startup. This would result in reloading
also failing even if a valid configuration file was put in place.
ASTERISK-23733 #close
Reported by: Richard Kenner
........
Merged revisions 429893 from http://svn.asterisk.org/svn/asterisk/branches/11
Richard Mudgett [Fri, 19 Dec 2014 17:31:28 +0000 (17:31 +0000)]
chan_dahdi: Don't ignore setvar when using configuration section scheme.
When the configuration section scheme of chan_dahdi.conf is used (keyword
dahdichan instead of channel) all setvar= options are completely ignored.
No variable defined this way appears in the created DAHDI channels.
* Move the clearing of setvar values to after the deferred processing of
dahdichan.
bridge: avoid leaking channel during blond transfer
After a blond transfer (start attended and hang up)
to a destination that also hangs up without answer,
the Local;1 channel was leaked and would show up on
core show channels. This was happening because the
attended state blond_nonfinal_enter() resetting the
props->transfer_target to null while releasing it's
own reference, which would later prevent props from
releasing another reference during destruction. The
change made here is simply to not assign the target
to NULL.
ASTERISK-24513 #close
Reported by: Mark Michelson
Review: https://reviewboard.asterisk.org/r/4262/
........
Merged revisions 429826 from http://svn.asterisk.org/svn/asterisk/branches/12
Kevin Harwell [Thu, 18 Dec 2014 15:50:26 +0000 (15:50 +0000)]
res_pjsip_sdp_rtp: wrong bridge chosen when the DTMF mode is not compatible
A native rtp bridge was being chosen (it shouldn't have been) when using two
pjsip channels with incompatible DTMF modes. This patch sets the rtp instance
property, AST_RTP_PROPERTY_DTMF, for the appropriate DTMF mode(s) for pjsip.
It was not being set before, meaning all DTMF modes for pjsip were being treated
as compatible, thus native bridging would be chosen as the bridge type when it
shouldn't have been.
Mark Michelson [Thu, 18 Dec 2014 15:34:16 +0000 (15:34 +0000)]
Prevent potential infinite outbound authentication loops in registration.
Prior to this patch, Asterisk would always respond to 401 responses to
registration attempts by trying to provide a registration with authentication
credentials. Even if subsequent attempts were rejected with 401 responses,
Asterisk would continue this behavior. If authentication credentials were
incorrect, this could continue forever.
With this patch, we keep track of whether we have attempted authentication
on an outbound registration attempt. If we already have, we don not try
again until the next attempt. This prevents the infinite loop scenario.
Mark Michelson [Thu, 18 Dec 2014 15:05:49 +0000 (15:05 +0000)]
Prevent possible race condition on dual redirect of channels in the same bridge.
The AST_FLAG_BRIDGE_DUAL_REDIRECT_WAIT flag was created to prevent bridges from
prematurely acting on orphaned channels in bridges. The problem with the AMI
redirect action was that it was setting this flag on channels based on the presence
of a PBX, not whether the channel was in a bridge. Whether a channel has a PBX
is irrelevant, so the condition has been altered to check if the channel is in a
bridge.
projects / thirdparty / asterisk.git / log
