When the stasis cache is used a hash is calculated for
retrieving or inserting messages. This change calculates
a hash when the message type is initialized that is then
used each time needed. This ensures that the hash is
calculated only once for the message type.
Change-Id: I4fe6bfdafb55bf5c322dd313fbd8c32cce73ef37
* Don't include pjlib.h twice in res_pjsip.h
* Consistently use #include <> form for pjproject includes.
(pjsip.h and pjlib.h)
Change-Id: I3f7b42044840de64edf7e9d7695cb60c45990dc7
Changing any Menuselect option in the `Compiler Flags` section causes a
full rebuild of the Asterisk source tree. Every enabled option causes
a #define to be added to buildopts.h, thus breaking ccache caching for
every source file that includes "asterisk.h". In most cases each option
only applies to one or two files. Now we only define those options for
the specific sources which use them, this causes much better cache
matching when working with multiple builds. For example testing code
with an without MALLOC_DEBUG will now use just over half the ccache
size, only main/astmm.o will have two builds cached instead of every
file.
Reorder main/Makefile so _ASTCFLAGS set on specific object files are all
together, sorted by filename. Stop adding -DMALLOC_DEBUG to CFLAGS of
bundled pjproject, this define is no longer used by any header so only
serves to break cache.
The only code change is a slight adjustment to how main/astmm.c is
initialized. Initialization functions always exist so main/asterisk.c
can call them unconditionally. Additionally rename the astmm
initialization functions so they are not exported.
Change-Id: Ie2085237a964f6e1e6fff55ed046e2afff83c027
In Solaris, the header <jansson.h> is in /usr/include/jansson. To find
Jansson even in such a subdirectory, the tool pkg-config is queried via
AST_PKG_CONFIG_CHECK. For those platforms, which do not list Jansson via
pkg-config, the previous check remains and is executed thereafter.
Because the check for the NetBSD Editline library uses the tool pkg-config
the code of PKG_PROG_PKG_CONFIG must be used. Because that check happens
earlier than Jansson, it must be placed in front of that.
ASTERISK-27991
Change-Id: I69ea0f379f87a50049654b2487c76ee1c04fa53a
When publishing a device state the change can be marked as being
cachable or not. If it is not cached the change is just published
to all interested and not stored away for later query. This was not
fully taken into account when publishing in stasis. The act of
publishing would create a topic for the device even if it may be
ephemeral.
This change makes it so messages which are not cached won't create
a topic for the device. If a topic does already exist it will be
published to but otherwise the change will only be published to
the device state all topic.
ASTERISK-27591
Change-Id: I18da0e8cbb18e79602e731020c46ba4101e59f0a
The "xmldoc dump" cli command was simply concatenating xml documents
into the output file. The resulting file had multiple "xml"
processing instructions and multiple root elements which is illegal.
Normally this isn't an issue because Asterisk has only 1 main xml
documentation file but codec_opus has its own file so if it's
downloaded and you do "xmldoc dump", the result is invalid.
* Added 2 new functions to xml.c:
ast_xml_copy_node_list creates a copy of a list of children.
ast_xml_add_child_list adds a list to an existing list.
* Modified handle_dump_docs to create a new output document and
add to it the children from each input file. It then dumps the
new document to the output file.
Change-Id: I3f182d38c75776aee76413dadd2d489d54a85c07
In the past there was an assertion in the ast_sched_del function
and in order to ensure it was useful the calling function name,
line number, and filename had to be passed in. This cause the ABI
to be different between dev mode and non-dev mode.
This assertion is no longer present so the special logic can be
removed to make it the same between them both.
Change-Id: Icbc69c801e357d7004efc5cf2ab936d9b83b6ab8
Support has been added for receiving a NACK request and handling it.
Now, Asterisk can detect when a NACK request should be sent and knows
how to construct one based on the packets we've received from the remote
end. A buffer has been added that will store out of order packets until
we receive the packet we are expecting. Then, these packets are handled
like normal and frames are queued to the core like normal. Asterisk
knows which packets to request in the NACK request using a vector
which stores the sequence numbers of the packets we are currently missing.
If a missing packet is received, cycle through the buffer until we reach
another packet we have not received yet. If the buffer reaches a certain
size, send a NACK request. If the buffer reaches its max size, queue all
frames to the core and wipe the buffer and vector.
According to RFC3711, the NACK request must be sent out in a compound
packet. All compound packets must start with a sender or receiver
report, so some work was done to refactor the current sender / receiver
code to allow it to be used without having to also include sdes
information and automatically send the report.
Also added additional functionality to ast_data_buffer, along with some
testing.
For more information, refer to the wiki page:
https://wiki.asterisk.org/wiki/display/AST/WebRTC+User+Experience+Improvements
ASTERISK-27810 #close
Change-Id: Idab644b08a1593659c92cda64132ccc203fe991d
* Merge the preload and load stages, use load ordering to try preload's
first. This fixes an issue where `preload=res_config_curl` would fail
unless res_curl and func_curl were also preloaded. Now it is only
required that those modules be loaded during startup: autoload or
regular load is good enough.
* The configuration option `require` and `preload-require` were only
effective if the modules failed to load. These options will now abort
Asterisk startup if required modules fail to reach the 'Running'
state.
* Missing or invalid 'module.conf' did not prevent startup. Asterisk
doesn't do anything without modules so this a fatal error.
Change-Id: Ie4176699133f0e3a823b43f90c3348677e43a5f3
Keep track if ICE candidates were in the SDP offer & only put them
in the corresponding SDP answer if the offer condaind ICE candidates
ASTERISK-27957 #close
Change-Id: Idf2597ee48e9a287e07aa4030bfa705430a13a92
A new option 'suppress_q850_reason_headers' has been added to the
endpoint object. Some devices can't accept multiple Reason headers and
get confused when both 'SIP' and 'Q.850' Reason headers are received.
This option allows the 'Q.850' Reason header to be suppressed.
The default value is 'no'.
ASTERISK-27949
Reported-by: Ross Beer
Change-Id: I54cf37a827d77de2079256bb3de7e90fa5e1deb1
The AMI action was directly sending the text to the channel driver.
However, this makes two threads attempt to handle media and runs afowl of
CHECK_BLOCKING.
* Queue a read action to make the channel's media handling thread actually
send the text message. This changes the AMI actions success/fail response
to just mean the text was queued to be sent not that the text actually got
sent. The channel driver may not even support sending text messages.
ASTERISK-27943
Change-Id: I9dce343d8fa634ba5a416a1326d8a6340f98c379
pjproject by default currently will follow media forked during an INVITE
on outbound calls if the To tag is different on a subsequent response as
that on an earlier response. We handle this correctly. There have
been reported cases where the To tag is the same but we still need to
follow the media. The pjproject patch in this commit adds the
capability to sip_inv and also adds the capability to control it at
runtime. The original "different tag" behavior was always controllable
at runtime but we never did anything with it and left it to default to
TRUE.
So, along with the pjproject patch, this commit adds options to both the
system and endpoint objects to control the two behaviors, and a small
logic change to session_inv_on_media_update in res_pjsip_session to
control the behavior at the endpoint level.
The default behavior for "different tags" remains the same at TRUE and
the default for "same tag" is FALSE.
Change-Id: I64d071942b79adb2f0a4e13137389b19404fe3d6
ASTERISK-27936
Reported-by: Ross Beer
* changes:
channel.c: Make CHECK_BLOCKING() save thread LWP id for messages.
channel.c: Fix usage of CHECK_BLOCKING()
autoservice: Don't start channel autoservice if the thread is a user interface.
There can be one and only one thread handling a channel's media at a time.
Otherwise, we don't know which thread is going to handle the media frames.
ASTERISK-27625
Change-Id: I4d6a2fe7386ea447ee199003bf8ad681cb30454e
There can be one and only one thread handling a channel's media at a time.
Otherwise, we don't know which thread is going to handle the media frames.
ASTERISK-27625
Change-Id: Ia341f1a6f4d54f2022261abec9021fe5b2eb4905
The CHECK_BLOCKING() macro is used to indicate if a channel's handling
thread is about to do a blocking operation (poll, read, or write) of
media. A few operations such as ast_queue_frame(), soft hangup, and
masquerades use the indication to wake up the blocked thread to reevaluate
what is going on.
ASTERISK-27625
Change-Id: I4dfc33e01e60627d962efa29d0a4244cf151a84d
Executing dialplan functions from either AMI or ARI by getting a variable
could place the channel into autoservice. However, these user interface
threads do not handle the channel's media so we wind up with two threads
attempting to handle the media.
There can be one and only one thread handling a channel's media at a time.
Otherwise, we don't know which thread is going to handle the media frames.
ASTERISK-27625
Change-Id: If2dc94ce15ddabf923ed1e2a65ea0ef56e013e49
It is invalid to typedef something more than once. Though not all gcc
compilers on different OS's complain about it.
Change-Id: I5a7d4565990c985822d61ce75bde0b45f9870540
Furthermore, allow OpenSSL configured with no-dh. Additionally, this change
allows auto-negotiation of the elliptic curve/group for servers, not only with
OpenSSL 1.0.2 but also with OpenSSL 1.1.0 and newer. This enables X25519
(since OpenSSL 1.1.0) and X448 (since OpenSSL 1.1.1) as a side-effect.
ASTERISK-27910
Change-Id: I5b0dd47c5194ee17f830f869d629d7ef212cf537
asterisk/tcptls.h was included (explicitly, implicitly, or transitively). Those
inclusions got replaced by forward declarations. As side effect, the inclusions
got completed.
ASTERISK-27878
Change-Id: I9d102728e30336d6522e5e4ae9e964013a0835f7
When RTP was originally created it had the ability to place a single
extension in an RTP packet. In practice people wanted to potentially
put multiple extensions in one and so RFC 5285 (obsoleted by RFC
8285) came into existence. This allows RTP extensions to be negotiated
with a unique identifier to be used in the RTP packet, allowing
multiple extensions to be present in the packet.
This change extends the RTP engine API to add support for this. A
user of it can enable extensions and the API provides the ability to
retrieve the information (to construct SDP for example) and to provide
negotiated information (from SDP). The end result is that the RTP
engine can then query to see if the extension has been negotiated and
what unique identifier is to be used. It is then up to the RTP engine
implementation to construct the packet appropriately.
The first extension to use this support is abs-send-time which is
defined in the REMB draft[1] and is a second timestamp placed in an
RTP packet which is for when the packet has left the sending system.
It is used to more accurately determine the available bandwidth.
ASTERISK-27831
[1] https://tools.ietf.org/html/draft-alvestrand-rmcat-remb-03
Change-Id: I508deac557867b1e27fc7339be890c8018171588
This function originally was used in chan_sip to enable some simplifying
assumptions and eventually was copy and pasted into res_pjsip_logger and
res_hep. Since it's replicated in three places, it's probably best to
move it into the public netsock2 API for these modules to use.
Change-Id: Id52e23be885601c51d70259f62de1a5e59d38d04
The stream topology has no lock of its own resulting in
another lock protecting it in some way (for example the
channel lock). If multiple channels are being juggled at
the same time this can be problematic. This change makes
the topology a reference counted object instead which
guarantees it will remain valid even without the channel
lock being held.
Change-Id: I4f4d3dd856a033ed55fe218c3a4fab364afedb03
How it works today:
media_cache tries to parse out the extension of the media file to be played
from the URI provided to Asterisk while caching the file.
What's expected:
Better will be to have Asterisk get extension from other ways too. One of the
common ways is to get the type of content from the CONTENT-TYPE header in the
HTTP response for fetching the media file using the URI provided.
Steps to Reproduce:
Provide a URL of the form: http://host/media/1234 to Asterisk for media
playback. It fails to play and logs show the following error line:
[Sep 15 15:48:05] WARNING [29148] [C-00000092] file.c:
File http://host/media/1234 does not exist in any format
Scenario this issue is blocking:
In the case where the media files are stored in some cloud object store,
following can block the media being played via Asterisk:
Cloud storage generally needs authenticated access to the storage. The way
to do that is by using signed URIs. With the signed URIs there's no way to
preserve the name of the file.
In most cases Cloud storage returns a key to access the object and preserving
file name is also not a thing there
ASTERISK-27286
Reporter: Gaurav Khurana
Change-Id: I1b14692a49b2c1ac67688f58757184122e92ba89
The OPTIONS support in PJSIP has organically grown, like many things in
Asterisk. It has been tweaked, changed, and adapted based on situations
run into. Unfortunately this has taken its toll. Configuration file
based objects have poor performance and even dynamic ones aren't that
great.
This change scraps the existing code and starts fresh with new eyes. It
leverages all of the APIs made available such as sorcery observers and
serializers to provide a better implementation.
1. The state of contacts, AORs, and endpoints relevant to the qualify
process is maintained. This state can be updated by external forces (such
as a device registering/unregistering) and also the reload process. This
state also includes the association between endpoints and AORs.
2. AORs are scheduled and not contacts. This reduces the amount of work
spent juggling scheduled items.
3. Manipulation of which AORs are being qualified and the endpoint states
all occur within a serializer to reduce the conflict that can occur with
multiple threads attempting to modify things.
4. Operations regarding an AOR use a serializer specific to that AOR.
5. AORs and endpoint state act as state compositors. They take input
from lower level objects (contacts feed AORs, AORs feed endpoint state)
and determine if a sufficient enough change has occurred to be fed further
up the chain.
6. Realtime is supported by using observers to know when a contact has
been registered. If state does not exist for the associated AOR then it
is retrieved and becomes active as appropriate.
The end result of all of this is best shown with a configuration file of
3000 endpoints each with an AOR that has a static contact. In the old
code it would take over a minute to load and use all 8 of my cores. This
new code takes 2-3 seconds and barely touches the CPU even while dealing
with all of the OPTIONS requests.
ASTERISK-26806
Change-Id: I6a5ebbfca9001dfe933eaeac4d3babd8d2e6f082