Richard Begg [Tue, 14 Mar 2017 21:22:42 +0000 (08:22 +1100)]
res_pjsip_sdp_rtp: RTP instance does not use same IP as explicit transport
Currently a wildcard address is used for the local RTP socket, which
will not always result in the same address as used by the SIP socket
(e.g. if explicit transport addresses are configured).
Use the transport's host address when binding new local RTP sockets if
available.
Mark Michelson [Tue, 7 Mar 2017 20:13:02 +0000 (14:13 -0600)]
Add rtcp-mux support
This commit adds support for RFC 5761: Multiplexing RTP Data and Control
Packets on a Single Port. Specifically, it enables the feature when
using chan_pjsip.
A new option, "rtcp_mux" has been added to endpoint configuration in
pjsip.conf. If set, then Asterisk will attempt to use rtcp-mux with
whatever it communicates with. Asterisk follows the rules set forth in
RFC 5761 with regards to falling back to standard RTCP behavior if the
far end does not indicate support for rtcp-mux.
The lion's share of the changes in this commit are in
res_rtp_asterisk.c. This is because it was pretty much hard wired to
have an RTP and an RTCP transport. The strategy used here is that when
rtcp-mux is enabled, the current RTCP transport and its trappings (such
as DTLS SSL session) are freed, and the RTCP session instead just
mooches off the RTP session. This leads to a lot of specialized if
statements throughout.
Sean Bright [Thu, 9 Mar 2017 17:05:12 +0000 (12:05 -0500)]
app_queue: Handle the caller being redirected out of a queue bridge
A caller can leave the Queue() application after being bridged with a
member in a few ways:
* Caller or member hangup
* Caller is transferred somewhere else (blind or atx)
* Caller is externally redirected elsewhere
The first 2 scenarios are currently handled by subscribing to stasis
messages, but the 3rd is not explicitly covered. If a caller is
redirected away from the Queue() application, the member who was last
bridged with that caller will remain in an "In use" state until the
caller hangs up.
This patch adds handling of the caller leaving the queue via
redirection. We monitor the caller-member bridge, and if the caller is
the one that leaves, we treat it the same as we would a caller hangup.
Matt Jordan [Tue, 14 Mar 2017 12:50:07 +0000 (07:50 -0500)]
res_pjsip_endpoint_identifier_ip: Add an option to match requests by header
This patch adds a new features to the endpoint identifier module,
'match_header'. When set, inbound requests are matched by a provided SIP
header: value pair. This option works in conjunction with the existing
'match' configuration option, such that if any 'match*' attribute
matches an inbound request, the request is associated with the specified
endpoint.
Since this module now identifies by more than just IP address,
appropriate renaming of the module and/or variables can be done in a
non-release branch.
Richard Mudgett [Tue, 14 Mar 2017 21:16:23 +0000 (16:16 -0500)]
pbx.c: Fix crash from malformed exten pattern.
Forgetting to indicate an exten is a pattern can cause a crash if the
"pattern" has a character set range. e.g., "9999[3-5]" The crash is due
to a buffer overwrite because the '-' exten eye-candy wasn't removed as
expected and overran the allocated space.
The buffer overwrite is fixed two ways in this patch.
1) Fix ext_strncpy() to distinguish between pattern and non-pattern
extens. Now '-' characters are removed when they are eye-candy and not
when they are part of a pattern character set. Since the function is
private to pbx.c, the return value now returns the number of bytes written
to the destination buffer instead of the strlen() of the final buffer so
the callers that care don't need to add one.
2) Fix callers to ext_strncpy() to supply the correct available buffer
size of the destination buffer.
Richard Begg [Tue, 14 Mar 2017 21:51:41 +0000 (08:51 +1100)]
chan_iax2: Reload of iax peer results in loss of host address/port
When using a non-dynamic peer address, build_peer() invalidates the
peer address structure by setting the address family to unspecified.
However, if dnsmgr is enabled, the subsequent call to ast_dnsmgr_lookup()
will not amend the peer address if the cache is still valid, resulting
in peer connectivity failures.
To fix this, we call ast_dnsmgr_refresh() instead.
Matt Jordan [Tue, 14 Mar 2017 20:12:28 +0000 (15:12 -0500)]
configure: Don't use the progress bar with curl when downloading to stdout
In some scenarios, such as when there may not be a terminal (such as
inside a Docker container), curl will apparently direct the progress bar
to stdout. This can cause extra data to be appended to a file curl'd
down to stdout, resulting in md5 verification failures.
This patch removes the progress bar, and tells curl to download the file
silently.
Matt Jordan [Tue, 14 Mar 2017 14:59:48 +0000 (09:59 -0500)]
funcs/func_devstate: Remove new line in Device field of during module load
During module loading of func_devstate, Asterisk emits the current
device state of all Custom device states currently stored in the AstDB.
This was erroneously including a new line character ('\n') to the end of
the device state, causing two new lines to be emitted in
DeviceStateChange AMI events.
Note that this only happened for those device state changes that
occurred during startup. Regular device state changes for Custom device
states are handled elsewhere, and did not have the newline.
ASTERISK-26643 #close
Reported by: Roman Bedros
Tested by: Matt Jordan
patches:
ami_devstate.diff uploaded by Roman Bedros (License 6842)
Matt Jordan [Tue, 14 Mar 2017 14:37:34 +0000 (09:37 -0500)]
main/stasis_cache: Demote the ERROR message when removing a nonexistent item
This patch demotes the ERROR message that is displayed when a
nonexistent item is removed from the Stasis cache. The genesis of this
demotion is due to chan_sip's realtime peers and their interaction with
Asterisk's core ast_endpoint code, but ostensibly it could happen from
other channel drivers as well.
Since Mark Michelson already did an excellent job of explaining on this
issue, it is quoted here for posterity:
"Internally, when a realtime peer is retrieved, Asterisk creates an
ast_endpoint structure. When that peer is destroyed, the ast_endpoint is
destroyed as well. Part of the destruction of the ast_endpoint involves
clearing the Stasis cache of all information about that endpoint. The
problem here is that the act of creating the ast_endpoint is not enough
to actually put any information in the Stasis cache. Instead, something
has to happen, such as a state change, in order for the Stasis cache to
have any information about that endpoint. When a device registers,
chan_sip creates an ast_endpoint structure, processes the REGISTER, and
then destroys the ast_endpoint. When the ast_endpoint is destroyed,
there is nothing to destroy in the Stasis cache, so an error message is
emitted. When you use rtcachefriends, ast_endpoint structures persist
for the lifetime of the module and so you do not see this error
message."
Joshua Colp [Sun, 12 Mar 2017 14:21:16 +0000 (14:21 +0000)]
chan_pjsip: Don't assume a session will have a channel.
When querying for PJSIP specific information using the dialplan
function CHANNEL() it is possible that the underlying session
will no longer have a channel associated with it. This is
most likely to occur when the RTCP HEP module attempts to get
the channel name. If this happens then a crash will occur.
This change just adds a check that the channel exists on the
session before querying it.
Mark Michelson [Tue, 7 Mar 2017 19:38:17 +0000 (13:38 -0600)]
res_http_websocket: Fix faulty read logic.
When doing some WebRTC testing, I found that the websocket would
disconnect whenever I attempted to place a call into Asterisk. After
looking into it, I pinpointed the problem to be due to the iostreams
change being merged in.
Under certain circumstances, a call to ast_iostream_read() can return a
negative value. However, in this circumstance, the websocket code was
treating this negative return as if it were a partial read from the
websocket. The expected length would get adjusted by this negative
value, resulting in the expected length being too large.
This patch simply adds an if check to be sure that we are only updating
the expected length of a read when the return from a read is positive.
Joshua Colp [Tue, 7 Mar 2017 11:22:18 +0000 (11:22 +0000)]
core: Add stream topology changing primitives with tests.
This change adds a few things to facilitate stream topology changing:
1. Control frame types have been added for use by the channel driver
to notify the application that the channel wants to change the stream
topology or that a stream topology change has been accepted. They are
also used by the indicate interface to the channel that the application
uses to indicate it wants to do the same.
2. Legacy behavior has been adopted in ast_read() such that if a
channel requests a stream topology change it is denied automatically
and the current stream topology is preserved if the application is
not capable of handling streams.
Tests have also been written which confirm the multistream and
non-multistream behavior.
Sean Bright [Mon, 6 Mar 2017 19:15:45 +0000 (14:15 -0500)]
pbx_spool: Gracefully handle long lines in call files
Per the linked issue, we aren't checking the buffer filled by fgets()
to determine if it contains a newline, so we will fail to correctly
parse the trailing portion of a long line.
This patch increases the buffer size from 256 to 1024, and skips any
line that exceeds that length, logging a warning in the process.
Richard Mudgett [Fri, 3 Mar 2017 03:27:00 +0000 (21:27 -0600)]
core: Cleanup ast_get_hint() usage.
* manager.c:manager_state_cb() Fix potential use of uninitialized hint[]
if a hint does not exist for the requested extension. Ran into this when
developing a testsuite test. The AMI event ExtensionStatus came out with
the hint header value containing garbage. The AMI event PresenceStatus
also had the same issue.
* manager.c:action_extensionstate() no need to completely initialize the
hint[]. Only initialize the first element.
Jørgen H [Thu, 16 Feb 2017 10:22:47 +0000 (10:22 +0000)]
res_pjsip WebRTC/websockets: Fix usage of WS vs WSS.
According to the RFC[1] WSS should only be used in the Via header
for secure Websockets.
* Use WSS in Via for secure transport.
* Only register one transport with the WS name because it would be
ambiguous. Outgoing requests may try to find the transport by name and
pjproject only finds the first one registered. This may mess up unsecure
websockets but the impact should be minimal. Firefox and Chrome do not
support anything other than secure websockets anymore.
* Added and updated some debug messages concerning websockets.
* security_events.c: Relax case restriction when determining security
transport type.
* The res_pjsip_nat module has been updated to not touch the transport
on Websocket originating messages.
George Joseph [Fri, 24 Feb 2017 21:30:33 +0000 (14:30 -0700)]
stream: Unit tests for stream read and tweaks framework
* Removed the AST_CHAN_TP_MULTISTREAM tech property. We now rely
on read_stream being set to indicate a multi stream channel.
* Added ast_channel_is_multistream convenience function.
* Fixed issue where stream and default_stream weren't being set on
a frame retrieved from the queue.
* Now testing for NULL being returned from the driver's read or
read_stream callback.
* Fixed issue where the dropnondefault code was crashing on a
NULL f.
* Now enforcing that if either read_stream or write_stream are
set when ast_channel_tech_set is called that BOTH are set.
* Added the unit tests.
Mark Michelson [Wed, 22 Feb 2017 21:11:29 +0000 (15:11 -0600)]
SDP: Add initial SDP state machine.
This introduces and documents the various states in the state machine.
This also introduces API functions that induce state changes, and places
TODO comments telling what needs to be done in addition to what is
already there. Those TODOs will be replaced with real code in upcoming
changes.
George Joseph [Tue, 28 Feb 2017 02:07:06 +0000 (19:07 -0700)]
res_pjsip_pubsub: Remove unneeded endpoint unref
When a subscription was being recreated and the endpoint wasn't
found, we were trying to unref the endpoint. This was causing
FRACKs. Removed the unref.
Jørgen H [Thu, 16 Feb 2017 10:16:47 +0000 (10:16 +0000)]
res_pjsip: Fix crash when contact has no status
This change fixes an assumption in res_pjsip that a contact will
always have a status. There is a race condition where this is
not true and would crash. The status will now be unknown when
this situation occurs.
George Joseph [Mon, 27 Feb 2017 18:25:01 +0000 (11:25 -0700)]
build: Warn if asterisk is installed in both 32 and 64 bit sys dirs
... and clean them both up on uninstall.
We've fixed the issue where 'make install' was installing to
/usr/lib on 64-bit systems that use /usr/lib64. Now we need
to clean up the remnants in /usr/lib.
* 'make install' now prints a warning if DESTDIR/ASTLIBDIR
contains 'lib64' and libasterisk* shared libraries or modules
are also found in DESTDIR/ASTLIBDIR with 'lib64' transformed
to 'lib'.
* 'make uninstall' ALWAYS cleans up both DESTDIR/ASTLIBDIR and
DESTDIR/ASTLIBDIR with 'lib64' transformed to 'lib'.
Joshua Colp [Mon, 27 Feb 2017 13:02:38 +0000 (13:02 +0000)]
bridge_native_rtp: Handle case where channel joins already suspended.
The bridge_native_rtp module did not properly handle the case where
a smart bridge operation occurs while a channel is suspended. In this
scenario the module would incorrectly set up local or remote RTP
bridging despite the media having to flow through Asterisk. The remote
endpoint would see two media streams and experience wonky audio.
The module has been changed so that it ensures both channels are
not suspended when performing the native RTP bridging and this
requirement has been documented in the bridge technology.
Joshua Colp [Fri, 24 Feb 2017 17:49:59 +0000 (17:49 +0000)]
config: Improve documentation and behavior of outbound_proxy option.
This change updates the documentation for the outbound_proxy option
to ensure it is consistently stated that a full SIP URI must be
provided for the option.
The res_pjsip_outbound_registration module has also been changed so
that the provided outbound_proxy value is checked to ensure it is a
URI and if not an error is output stating so.
Joshua Colp [Thu, 23 Feb 2017 19:03:15 +0000 (19:03 +0000)]
channel: Add ast_read_stream function for reading frames from all streams.
This change introduces an ast_read_stream function and callback in
the channel technology which allows reading frames from all streams
and not just the default streams.
The stream number has also been added to frames. This is to allow the
case where frames are queued onto the channel instead of being read
directly from the driver.
This change does impose a restriction on reading though: a chain of
frames can only contain frames from the same stream.
George Joseph [Fri, 10 Feb 2017 00:05:01 +0000 (17:05 -0700)]
pjproject_bundled: Update for pjproject 2.6
* Removed all 2.5.5 functional patches.
* Updated usages of pj_release_pool to be "safe".
* Updated configure options to disable webrtc.
* Updated config_site.h to disable webrtc in pjmedia.
* Added Richard Mudgett's recent resolver patches.
George Joseph [Thu, 23 Feb 2017 21:49:17 +0000 (14:49 -0700)]
build: Execute ldconfig to build cache. (take two)
On some platforms a multiarch approach is used for libraries.
The build system does not take this into account and still
places libraries into the lib directory if no --libdir is
specified to configure. On initial startup this results in
libasteriskssl.so not being found, as it is not in the multiarch
lib directory. To make matters worse, options were being passed
to ldconfig on both Linux and FreeBSD that actually prevented
the rebuild of the cache.
* Fedora has a /usr/share/config.site that automatically tells
autoconf to use /usr/lib64 but CentOS does not. This logic was
copied to configure.ac and modified so systems like Ubuntu,
which still use /usr/lib for 64-bit systems, aren't affected.
Now that we have them in the correct directory...
In order for the system loader to find libasteriskssl and
libasteriskpj, one of 3 things has to happen...
- The linker cache must be rebuilt including the directory
where the libasterisk* libraries were installed. Only root
can rebuild the cache. This was busted.
- We have to link the asterisk binary with an rpath pointing
to the directrory where the libasterisk* libraries were
installed. This makes things very complicated and will happen
over the collective dead bodies of everyone who's had to
package a distribution with an rpath.
- Finally, you can start asterisk with LD_LIBRARY_PATH set to the
directrory where the libasterisk* libraries were installed.
There are no other options. So...
* The invokation of ldconfig has been moved from main/Makefile
to ASTTOPDIR/Makefile, the options have been removed, and
DESTDIR/ASTLIBDIR appended. If you aren't root, you will be
warned after the "Asterisk Installation Compete" banner that
you must re-run 'make install' as root, manually run
'ldconfig DESTDIR/ASTLIBDIR' as root, or run asterisk with
LD_LIBRARY_PATH.
Joshua Colp [Wed, 22 Feb 2017 11:00:57 +0000 (11:00 +0000)]
channel: Add support for writing to a specific stream.
This change adds an ast_write_stream function which allows
writing a frame to a specific media stream. It also moves
ast_write() to using this underneath by writing media
frames provided to it to the default streams of the channel.
Existing functionality (such as audiohooks, framehooks, etc)
are limited to being applied to the default stream only.
Unit tests have also been added which test the behavior of
both non-multistream and multistream channels to confirm that
the write() and write_stream() callbacks are invoked
appropriately.
frahaase [Fri, 12 Aug 2016 16:23:18 +0000 (18:23 +0200)]
Binaural synthesis (confbridge): Adds binaural synthesis to bridge_softmix.
Adds binaural synthesis to bridge_softmix (via convolution using libfftw3).
Binaural synthesis is conducted at 48kHz.
For a conference, only one spatial representation is rendered.
The default rendering is applied for mono-capable channels.
Sean Bright [Wed, 22 Feb 2017 14:53:25 +0000 (09:53 -0500)]
res_config_ldap: Various code improvements
The initial motivation for this patch was to properly handle memory
allocation failures - we weren't checking the return values from the
various LDAP library allocation functions.
In the process, because update_ldap() and update2_ldap() were
substantially the same code, they've been consolidated.
Michael L. Young [Wed, 22 Feb 2017 19:08:05 +0000 (14:08 -0500)]
build_tools: Fix download_externals to allow the use of curl or wget
Not sure if this is really a bug versus an improvement. I can see it being
viewed as a bug though by some.
The current build_tools/download_externals file depends on wget in order to
download external modules. The current build system is able to discover
which tool to use for fetching remote files - either wget or curl.
This patch takes advantage of this capability by modifying the two calls to
the wget binary to instead use what was discovered by the build system.
The patterns provided by pbx_realtime were checked in the order in
which they were returned from the realtime backend. If there was
overlap between multiple patterns, the first one to correctly match was
chosen even though it may not have been the best match.
We now sort the patterns descending by their length and compare in that
order. There may be cases where this still results in a sub-optimal
match, but this patch should improve the overall behavior.
ASTERISK-18271 #close
Reported by: Charlie Smurthwaite
projects / thirdparty / asterisk.git / log
