Sungtae Kim [Wed, 9 Jan 2019 10:27:03 +0000 (10:27 +0000)]
http.c: Support separated HTTP request
Currently, the Asterisk does not support seperated HTTP request.
This patch make the Asterisk enables to wait lest part of HTTP request.
Also increases acceptable HTTP body length to 40k to support more
larger request.
George Joseph [Wed, 20 Feb 2019 18:48:25 +0000 (11:48 -0700)]
Core: Increase AST_PBX_MAX_STACK to 512 if not LOW_MEMORY
The current settings AST_PBX_MAX_STACK is 128 entries which is
too low for some FreePBX installations with complex parking
arrangements. Increased to 512 if LOW_MEMORY is not defined.
Joshua C. Colp [Wed, 20 Feb 2019 18:22:31 +0000 (14:22 -0400)]
stasis: Store subscriber uniqueids with topic statistics.
This change provides an easier mechanism to determine which
subscribers are subscribed to a topic. Using this you can
inspect the specific subscribers for further details.
George Joseph [Fri, 15 Feb 2019 18:53:50 +0000 (11:53 -0700)]
taskprocessor: Enable subsystems and overload by subsystem
To prevent one subsystem's taskprocessors from causing others
to stall, new capabilities have been added to taskprocessors.
* Any taskprocessor name that has a '/' will have the part
before the '/' saved as its "subsystem".
Examples:
"sorcery/acl-0000006a" and "sorcery/aor-00000019"
will be grouped to subsystem "sorcery".
"pjsip/distributor-00000025" and "pjsip/distributor-00000026"
will bn grouped to subsystem "pjsip".
Taskprocessors with no '/' have an empty subsystem.
* When a taskprocessor enters high-water alert status and it
has a non-empty subsystem, the subsystem alert count will
be incremented.
* When a taskprocessor leaves high-water alert status and it
has a non-empty subsystem, the subsystem alert count will be
decremented.
* A new api ast_taskprocessor_get_subsystem_alert() has been
added that returns the number of taskprocessors in alert for
the subsystem.
* A new CLI command "core show taskprocessor alerted subsystems"
has been added.
* A new unit test was addded.
REMINDER: The taskprocessor code itself doesn't take any action
based on high-water alerts or overloading. It's up to taskprocessor
users to check and take action themselves. Currently only the pjsip
distributor does this.
* A new pjsip/global option "taskprocessor_overload_trigger"
has been added that allows the user to select the trigger
mechanism the distributor uses to pause accepting new requests.
"none": Don't pause on any overload condition.
"global": Pause on ANY taskprocessor overload (the default and
current behavior)
"pjsip_only": Pause only on pjsip taskprocessor overloads.
* The core pjsip pool was renamed from "SIP" to "pjsip" so it can
be properly grouped into the "pjsip" subsystem.
* stasis taskprocessor names were changed to "stasis" as the
subsystem.
* Sorcery core taskprocessor names were changed to "sorcery" to
match the object taskprocessors.
Kevin Harwell [Fri, 8 Feb 2019 19:07:13 +0000 (13:07 -0600)]
ARI event type filtering
Event type filtering is now enabled, and configurable per application. An app is
now able to specify which events are sent to the application by configuring an
allowed and/or disallowed list(s). This can be done by issuing the following:
PUT /applications/{applicationName}/eventFilter
And then enumerating the allowed/disallowed event types as a body parameter.
sungtae kim [Thu, 14 Feb 2019 23:09:30 +0000 (00:09 +0100)]
chan_pjsip: Changed to continued after invalid media for pjsip show channelstats
Currently, the pjsip show channelstats cli does not show channel's
stats after hits the invalid channel info. This makes hard to use
this cli. Changed to keep iterate after hits the invalid channel
info.
Joshua Colp [Tue, 19 Feb 2019 16:06:32 +0000 (16:06 +0000)]
CI: Use tmpfs option to Docker instead of mount.
Some tests require Asterisk to execute scripts which
are stored in /tmp. When mount is used for tmpfs there
is no ability to allow scripts to be executed from
that location.
This change switches to using tmpfs which can be told
to allow executables to be run from /tmp.
Torrey Searle [Tue, 12 Feb 2019 09:50:55 +0000 (10:50 +0100)]
res/res_rtp_asterisk: clear smoother when local bridging
p2p_write updates txformat but doesn't require a smoother. If a smoother
was created by another bridge type the smoother could fall out of date causing
one way audio issues. To prevent this the smoother is now destroyed on the
start of native bridge.
Joshua Colp [Thu, 7 Feb 2019 15:52:56 +0000 (15:52 +0000)]
ci: Rerun unit tests when non-code changes occur.
This change makes it so that even if non-code changes
occur (such as commit message changing) unit tests
will still be run and result in a verification.
Kevin Harwell [Thu, 31 Jan 2019 19:29:05 +0000 (13:29 -0600)]
res_pjsip_registrar: lock transport monitor when setting 'removing' flag
A previous patch attempt to mitigate blocked threads on transport shutdown for
a given contact. It was thought that a second lock could be avoided by checking
the 'removing' flag on the transport monitor twice (once before and once after
the normal named aor locking). However as with usual threading issues if the
timing was right the original problem still occured.
This patch adds locking around the first 'removing' flag check and set, thus
nullifying the secondary check, so it was removed.
Joshua Colp [Wed, 6 Feb 2019 12:16:01 +0000 (12:16 +0000)]
res_odbc: Add basic query logging.
When Asterisk is connected and used with a database the response
time of the database can cause problems in Asterisk if it is long.
Normally the only way to see this problem would be to retrieve a
backtrace from Asterisk and examine where things are blocked, or
examine the database to see if there is any indication of a
problem.
This change adds some basic query logging to make it easier to
investigate such a problem. When logging is enabled res_odbc will
now keep track of the number of queries executed, as well as the
query that has taken the longest time to execute. There is also
an option which will cause a WARNING message to be output if a
query takes longer than a configurable amount of time to execute.
This makes it easier and clearer for users that their database may
be experiencing a problem that could impact Asterisk.
George Joseph [Mon, 4 Feb 2019 13:09:57 +0000 (06:09 -0700)]
bundled-jansson: On OpenSuse Leap libjansson.a was placed in lib64
On OpenSuse Leap, libjansson.a is installed in
third-party/jansson/dest/lib64 instead of lib (which is where
the top-level makeopts looks). This causes a link failure.
* Updated jansson/Makefile to add an explicit --libdir to force
the installation to third-party/jansson/dest/lib.
Ben Ford [Tue, 29 Jan 2019 16:48:49 +0000 (10:48 -0600)]
res_stasis: Auto-create context and extens on Stasis app launch.
At AstriCon, there was a strong desire for the ability to completely
bypass dialplan when using ARI. This is possible through the automatic
creation of a context and a couple of extensions whenever an application
is started.
For example, if you have an application named 'ari-example', a context
named 'stasis-ari-example' will be automatically created whenever this
application is started as long as one does not already exist. Two
extensions (a match-all extension for Stasis and a 'h' extension) are
created within this context. Any endpoint that registers to Asterisk
within this context will send all calls to the corresponding Stasis
application. When the application is destroyed, the context is removed.
Kevin Harwell [Tue, 15 Jan 2019 23:20:30 +0000 (17:20 -0600)]
pjsip/config_global: regcontext context not created
The context specified by 'regcontext' was not being created, so when Asterisk
attempted to later dynamically add an extension it would fail. This patch now
creates the context if a 'regcontext' is specified.
George Joseph [Tue, 22 Jan 2019 15:02:06 +0000 (08:02 -0700)]
media_index.c: Refactored so it doesn't cache the index
Testing revealed that the cache added no benefit but that it could
consume excessive memory.
Two new index related functions were created:
ast_sounds_get_index_for_file() and ast_media_index_update_for_file()
which restrict index updating to specific sound files.
The original ast_sounds_get_index() and ast_media_index_update()
calls are still available but since they no longer cache the results
internally, developers should re-use an index they may already have
instead of calling ast_sounds_get_index() repeatedly. If information
for only a single file is needed, ast_sounds_get_index_for_file()
should be called instead of ast_sounds_get_index().
The media_index directory scan code was elimininated in favor of
using the existing ast_file_read_dirs() function.
Since there's no more cache, ast_sounds_index_init now only
registers the sounds cli commands instead of generating the
initial index and subscribing to stasis format register/unregister
messages.
ast_sounds_reindex() is now a no-op but left for backwards
compatibility.
loader.c no longer registers "sounds" as a special reload target.
Both the sounds cli commands and the sounds ari resources were
refactored to only call ast_sounds_get_index() once per invocation
and to use ast_sounds_get_index_for_file() when a specific sound
file is requested.
Fix deadlock handling subscribe req during res_parking reload
Split destroy_hint method to separate hint removal and extension hint
state changed callback, the latter now called via stasis.
This avoids deadlock between res_parking reload that is removing the
parking lot and the related hint and subscribe requests coming for the
same parking lot.
eyalhasson [Tue, 22 Jan 2019 15:24:23 +0000 (17:24 +0200)]
format_g726: add support for seeking
Added support for the seek function in format_g726
so playback can start from anywhere.
Before the fix, playback of g726 files
always started from the beginning.
Paulo Vicentini [Thu, 8 Nov 2018 10:21:03 +0000 (11:21 +0100)]
res/res_pjsip: Fix crash due to misuse of session->media between threads.
This patch makes sure that thread running ast_taskprocessor_execute
cannot suddenly dispose the session->media object making the other
threads (running pbx_thread / bridge_channel_ind_thread) crash when they
try to access the pointer to invalid memory. We were experiencing a crash due
to a misuse of session->media container between threads running
(bridge_channel_ind_thread/pbx_thread) and the thread running
ast_taskprocessor_execute. Depending on the SIP flow (during a disconnection)
and the threads' code path, the session->media container was being destroyed
(and set to NULL) by the thread running ast_taskprocessor_execute while the
thread running t38_framehook_read was still referring to it.
Now res_pjsip_t38 is referring a session_media in a datastore.
Jeremy Lainé [Wed, 23 Jan 2019 10:45:56 +0000 (11:45 +0100)]
res_http_websocket: ensure control frames do not interfere with data
Control frames (PING / PONG / CLOSE) can be received in the middle of a
fragmented message. In order to ensure they do not interfere with the
reassembly buffer, we exit early and do not return the payload to the
caller.
Gerald Schnabel [Tue, 22 Jan 2019 21:03:22 +0000 (22:03 +0100)]
manager_channels: Fix throwing of HangupHandler manager events
The type value extracted from stasis message data in channel_hangup_handler_cb
isn't compared against the valid values "run", "pop" and "push". Thus the
manager events HangupHandlerPush, HangupHandlerPop and HangupHandlerRun are
never thrown.
Chris-Savinovich [Tue, 22 Jan 2019 17:48:57 +0000 (11:48 -0600)]
Test_cel: Fails when DONT_OPTIMIZE is off
A bug in GCC causes TEST_CEL to return
failure under the following conditions:
1. TEST_FRAMEWORK on
2. DONT_OPTIMIZE off
3. Fedora and Ubuntu
4. GCC 8.2.1
5. Test name: test_cel_dial_pickup
6. There must exist a certain combination of multithreading.
The bug affects arithmetic calculations when the optimization level
is bigger than O1 and the -fpartial-inline flag is on. Provided these
conditions, function ast_str_to_lower() fails to convert to lower case
due to said function being of type force_inline. The solution is to
remove the "force_inline" type declaration from function ast_str_to_lower()
Change-Id: Ied32e0071f12ed9d5f3b4cdd878b2532a1c769d7
George Joseph [Mon, 10 Dec 2018 13:20:06 +0000 (06:20 -0700)]
app_voicemail: Add Mailbox Aliases
You can now define an "aliases" context in voicemail.conf
whose entries point to actual mailboxes. These can be used anywhere
the mailbox is specified.
Example:
[general]
aliasescontext = myaliases
[default]
1234 = yadayada
[myaliases]
4321@devices = 1234@default
Now you can use 4321@devices to refer to the 1234@default mailbox.
This can be useful to provide channel drivers with constant
mailbox specifications such as <extension>@devices leaving
app_voicemail to control exactly which mailbox the alias points to.
Now, only voicemail has to be reloaded to make changes instead of
individual channel drivers which are usually more expensive to
reload.
Kevin Harwell [Tue, 22 Jan 2019 17:48:41 +0000 (11:48 -0600)]
res_pjsip_registrar: mitigate blocked threads on reliable transport shutdown
When a reliable transport is shutdown it's possible for the pjsip registrar
resource shutdown handler to get called multiple times. If this happens and one
of the threads is taking "too long" (slow database call for instance) then the
others get blocked waiting to delete.
Since it only takes one to delete the contact then the other threads should be
able to continue on if one of the threads is currently "deleting". This patch
makes it so now when a thread enters the shutdown handler it checks to see if a
thread is currently already "deleting". If so, then the thread does not attempt
to get the lock, and instead continues on thus avoiding the blockage.
George Joseph [Tue, 22 Jan 2019 15:02:37 +0000 (08:02 -0700)]
pjproject_bundled: Add patch for double free issue in timer heap
Fixed #2172: Avoid double reference counter decrements in
timer in the scenario of race condition between
pj_timer_heap_cancel() and pj_timer_heap_poll().
Jeremy Lainé [Tue, 8 Jan 2019 07:38:41 +0000 (08:38 +0100)]
res_http_websocket: respond to CLOSE opcode
This ensures that Asterisk responds properly to frames received from a
client with opcode 8 (CLOSE) by echoing back the status code in its own
CLOSE frame.
Handling of the CLOSE opcode is moved up with the rest of the opcodes so
that unmasking gets applied. The payload is no longer returned to the
caller, but neither ARI nor the chan_sip nor pjsip made use of the
payload, which is a good thing since it was masked.
Sean Bright [Fri, 18 Jan 2019 22:11:18 +0000 (17:11 -0500)]
pjsip_transport_management: Shutdown transport immediately on disconnect
The transport management code that checks for idle connections keeps a
reference to PJSIP's transport for IDLE_TIMEOUT milliseconds (32000 by
default). Because of this, if the transport is closed before this
timeout, the idle checking code will keep the transport from actually
being shutdown until the timeout expires.
Rather than passing the AO2 object to the scheduler task, we just pass
its key and look it up when it is time to potentially close the idle
connection. The other transport management code handles cleaning up
everything else for us.
Additionally, because we use the address of the transport when
generating its name, we concatenate an incrementing ID to the end of the
name to guarantee uniqueness.
Valentin Vidic [Sun, 20 Jan 2019 18:15:51 +0000 (19:15 +0100)]
channel.c: Fix segfault with Monitor(wav,file,i)
If the Monitor is started with the i option the read_stream will be
NULL. One code path in channel.c checks if write_stream is set but than
uses read_stream instead causing a segfault.
Joshua C. Colp [Thu, 10 Jan 2019 19:34:32 +0000 (15:34 -0400)]
stasis / manager / ari: Better filter messages.
Previously both AMI and ARI used a default route on
their stasis message router to handle some of the
messages for publishing out their respective
connection. This caused messages to be given to
their subscription that could not be formatted
into AMI or JSON.
This change adds an API call to the stasis message
router which allows a default route to be set as well
as formatters that the default route is expecting.
This allows both AMI and ARI to specify that their
default route only wants messages of their given
formatter. By doing so stasis can more intelligently
filter at publishing time so that they do not receive
messages which will not be turned into AMI or JSON.
Sean Bright [Fri, 4 Jan 2019 23:14:45 +0000 (18:14 -0500)]
res_pjsip_transport_websocket: Don't assert on 0 length payloads
If an external pjproject is used and built with assertions enabled, we
will assert if passed a payload length of 0, so treat empty frames as if
we didn't receive them.
Alexei Gradinari [Fri, 11 Jan 2019 16:34:08 +0000 (11:34 -0500)]
res_pjsip: add option to disable ContactStatus event when contact is updated
This patch adds a new PJSIP global configuration option
'send_contact_status_on_update_registration' to be able to have the same
performance benefits as version 16.
By default old behavior, i.e. the ContactStatus event will be sent when a
device refreshes its registration.
mohitdhiman [Mon, 7 Jan 2019 18:04:43 +0000 (23:34 +0530)]
stasis/endpoint: Fix memory leak of channel_ids in ast_endpoint structure.
During Bridging of two channels if masquerade operation is performed on a
channel (clone channel) which was created with endpoint details
(ast_channel_alloc_with_endpoint()) and the original channel which is created
without endpoint details (ast_channel_alloc()) then both the channels must
exchange their endpoint details or else after masquerade when clone channel
is being destroyed the endpoint cleanup callbacks will be destroyed too and
after call completion unique_id of original channel will still be there in
ast_endpoint structure's channel_ids container.
This prevents use-after-scope issues when unwinding the stack,
which happens in reverse order. The varname variable needs to
remain alive for the destruction to be able to access it.
Issue was found using clang + address-sanitizer.
Bryan Boatright [Wed, 2 Jan 2019 17:44:41 +0000 (11:44 -0600)]
app_voicemail: Fix Channel variable VM_MESSAGEFILE for "urgent" voicemail
If a voicemail is marked "urgent" then the VM_MESSAGEFILE channel variable is
not updated correctly since urgent messages are in a different directory. The
fix is to update the channel variable when the path to the urgent message is
created.
Richard Mudgett [Wed, 19 Dec 2018 19:02:35 +0000 (13:02 -0600)]
stasic.c: Fix printf format type mismatches with arguments.
An int64_t is not likely the same size as a long.
* Changed the int64_t values in the statistics structs to longs so casting
is not necessary when generating the formatted CLI output. The offending
members did not need to be int64_t anyway as they were only set by an int
type variable which was already truncating bits.
* Reordered the statistics structs to reduce potential padding bytes.
George Joseph [Mon, 24 Dec 2018 17:42:36 +0000 (10:42 -0700)]
ast_coredumper: Refactor the pid determination process
In order to get a dump of the running process, we need to find the
pid of the main asterisk process. This can be tricky if there are
also instances of "asterisk -r" running or if an alternate location
for asterisk.conf was specified on the command line with the -C
option that also specified an alternation location for the pid file.
So now...
1. We find the asterisk executable with "which" or the --asterisk-bin
command line option.
2. If there's only 1 process with an executable path that matches,
we use that pid. If not...
3. We try "<asterisk-bin> -rx 'core show settings'" and parse the
output to find the pidfile, then read that for the pid. If that
didn't work...
4. We get a list of all the pids matching <asterisk-bin> and look
in /proc/<pid>/cmdline for a -C argument and retry the "core show
settings" using the same -C option. We can't parse the output
of "ps" to get the -C path because it may contain spaces. The
contents of /proc/<pid>/cmdline are delimited by NULLs. For BSDs
we may have to mount /proc first. :(
Alexei Gradinari [Wed, 19 Dec 2018 21:27:12 +0000 (16:27 -0500)]
RTP: reset DTMF last seqno/timestamp on RTP renegotiation
The remote side may start a new stream when renegotiating RTP.
Need to reset the DTMF last sequence number and the timestamp
of the last END packet on RTP renegotiation.
If the new time stamp is lower then the timestamp of the last DTMF END packet
the asterisk drops all DTMF frames as out of order.
This bug was caught using Cisco ip-phone SPA5XX and codec g722.
On SIP session update the SPA50X resets stream and a new timestamp is twice
smaller then the previous.
Richard Mudgett [Wed, 19 Dec 2018 18:39:08 +0000 (12:39 -0600)]
backtrace.c: Fix casting pointer to/from integral type.
The backtrace library bfd.h include file does not get the sizes of
pointers and ints right on some platforms. On my old test box the size
of bfd_vma is 8 while the size of a pointer is 4. gcc on the box
complains of the integer casting to/from pointers size mismatch.
* uintptr_t to the rescue by doing an appropriate two stage cast.
George Joseph [Tue, 18 Dec 2018 16:33:50 +0000 (09:33 -0700)]
app_voicemail: Don't delete mailbox state unless mailbox is deleted
The free_user function was automatically deleting the stasis mailbox
state but this only makes sense when the mailbox is actually
deleted, not just the structure freed. This was causing issues
where leave_voicemail would publish the mwi message to stasis and
delete the state before the message could be processed by
res_pjsip_mwi.
* Removed the delete of state from free_user().
* Created a new free_user_final() function that both frees the data
structure and deletes the state. This function is only called
during module load/unload where it's appropriate to delete the
state.
Sean Bright [Thu, 13 Dec 2018 21:56:50 +0000 (16:56 -0500)]
res_format_attr_h264.c: Make sure profile-level-id fmtp attribute is set
The profile-iop octet (the 2nd) of profile-level-id can be zero
according to RFC 6184 Section 8.1. So we ignore its value when deciding
to include profile-level-id in the outgoing SDP.
projects / thirdparty / asterisk.git / log
