Corey Farrell [Sat, 10 Mar 2018 09:33:33 +0000 (04:33 -0500)]
core: Remove non-critical cleanup from startup aborts.
When built-in components of Asterisk fail to start they cause the
Asterisk startup to abort. In these cases only the most critical
cleanup should be performed - closing databases and terminating
proceses. These cleanups are registered using ast_register_atexit, all
other cleanups should not be run during startup abort.
The main reason for this change is that these cleanup procedures are
untestable from the partially initialized states, if they fail it could
prevent us from ever running the critical cleanup with ast_run_atexits.
Create separate initialization for dns_core.c to be run unconditionally
during startup instead of being initialized by the first dns resolver to
be registered. This ensures that 'sched' is initialized before it can be
potentially used.
Replace ast_register_atexit with ast_register_cleanup in media_cache.c.
There is no reason for this cleanup to happen unconditionally.
Alexander Traud [Mon, 12 Mar 2018 11:40:11 +0000 (12:40 +0100)]
install_prereq: Update FreeBSD libraries.
Because the code review system Gerrit creates merge conflicts even when one line
apart another change happened, the previous update to the FreeBSD libraries had
to be rebased via Git. Because of a break for training of the original
contributor, this rebase was done by another contributor and the variant for
Asterisk 13 was cherry-picked to all branches. By this, dependencies for new
features added in newer Asterisk version got lost. This can be seen, when not
the original path set but a previous patch set is compared.
This change here fixes this by adding those (optional) dependencies for
Asterisk 15 and newer (again).
Alexander Traud [Mon, 12 Mar 2018 09:11:15 +0000 (10:11 +0100)]
res_srtp: Add support for libsrtp2.x on openSUSE.
Since ASTERISK-27253, no symbols from the header srtp2/crypto_types.h are used
anymore. Therefore, its include statement can be removed. This allows to compile
Asterisk on platforms which do not offer this private header, like openSUSE.
Alexander Traud [Fri, 9 Mar 2018 09:23:53 +0000 (10:23 +0100)]
BuildSystem: For consistency, avoid double-checking via if clauses.
In the script ./configure, AST_EXT_LIB_CHECK and AST_PKG_CONFIG_CHECK first test
whether parameter 1 was already found. Consequently, an if-test on PBX_ just a
line below is redundant, if exactly the same parameter 1 is used again.
No performance gain is expected by this change. However, because this strategy
is used all over in ./configure except for two places, this change aims to
create more consistency: Only do something different if there is a reason to do
so.
Alexander Traud [Tue, 6 Mar 2018 12:28:15 +0000 (13:28 +0100)]
BuildSystem: Enable PortAudio in NetBSD.
In NetBSD, PortAudio 1 is still the default version. PortAudio 2 can be
installed side by side but gets placed in a 'portaudio2' subdirectory. To
find PortAudio 2 even in a subdirectory, the tool pkg-config is queried via
AST_PKG_CONFIG_CHECK. For those platforms, which do not list PowerAudio 2
via pkg-config, the previous check remains and is executed thereafter.
Corey Farrell [Wed, 7 Mar 2018 20:36:17 +0000 (15:36 -0500)]
Replace direct checks of option_debug with DEBUG_ATLEAST macro.
Checking option_debug directly is incorrect as it ignores file/module
specific debug settings. This system-wide change replaces nearly all
direct checks for option_debug with the DEBUG_ATLEAST macro.
Richard Mudgett [Wed, 7 Mar 2018 19:13:42 +0000 (13:13 -0600)]
BuildSystem regression: Fix errors reported by clean targets.
Doing a 'make clean', 'make distclean', or 'make dist-clean' gets errors
about an invalid shell option: "/bin/sh: 0: Illegal option -".
The clean targets do not include the makeopts file which defines GREP and
LDCONFIG because the file may not exist and the distclean/dist-clean
targets will delete it anyway.
Alexander Traud [Wed, 7 Mar 2018 15:32:05 +0000 (16:32 +0100)]
utils: In Solaris, avoid a warning about an unused variable.
When HAVE_GETHOSTBYNAME_R_5 was set by the script ./configure, GCC 7.3.0 found
an unused variable. Actually, the variable was used (set to a dummy value) but
the compiler optimization might have removed that. Instead, this change ensures
that the variable 'res' is only used when it is really required.
lvl [Tue, 20 Feb 2018 17:48:50 +0000 (18:48 +0100)]
res_pjsip_session: properly handle SDP from a forked call with early media
In handle_negotiated_sdp(), use session->active_media_state when
session->pending_media_state is empty. The 200's SDP should be fed into
handle_negotiated_sdp_session_media() together with the already negotiated
state, which is now in session->active_media_state instead. Only if both
the session's pending and active media are empty should
handle_negotiated_sdp() abort.
Richard Mudgett [Tue, 27 Feb 2018 21:40:18 +0000 (15:40 -0600)]
pjproject: Add cache_pools debugging option.
The pool cache gets in the way of finding use after free errors of memory
pool contents. Tools like valgrind and MALLOC_DEBUG don't know when a
pool is released because it gets put into the cache instead of being
freed.
* Added the "cache_pools" option to pjproject.conf. Disabling the option
helps track down pool content mismanagement when using valgrind or
MALLOC_DEBUG. The cache gets in the way of determining if the pool
contents are used after free and who freed it.
To disable the pool caching simply disable the cache_pools option in
pjproject.conf and restart Asterisk.
Michael Cargile [Sat, 24 Feb 2018 03:24:10 +0000 (22:24 -0500)]
apps/app_amd.c: Fixed total time and silence calculations
Between Asterisk 11 and Asterisk 13 there was a significant increase
in the number of AST_FRAME_NULL frames being processed by app_amd.c's
main loop. Each AST_FRAME_NULL frame was being counted as 100ms
towards the total time and silence. This may have been accurate
when app_amd.c was orginally added, but it is not in Asterisk 13.
As such the total analysis time and silence calculations were way
off effectively breaking app_amd.c
* Additional debug messages were added
* AST_FRAME_NULL are now ignored
Ben Ford [Thu, 22 Feb 2018 20:27:26 +0000 (14:27 -0600)]
Add extended properties to rtp_engine for RTP retransmission support.
A couple of additional properties are needed in rtp_engine to enable
support for packet retransmission: AST_RTP_PROPERTY_RETRANS_RECV and
AST_RTP_PROPERTY_RETRANS_SEND. These will both be enabled automatically
if an endpoint has the webrtc option enabled. While this adds no
functionality currently, it will serve as a building block for future
changes for RTP retransmission support.
For more information, refer to the wiki page:
https://wiki.asterisk.org/wiki/display/AST/WebRTC+User+Experience+Improvements
Corey Farrell [Fri, 23 Feb 2018 16:09:46 +0000 (11:09 -0500)]
core: Fix handling of maximum length lines in config files.
When a line is the maximum length "\n" is found at sizeof(buf) - 2 since
the last character is actually the null terminator. In addition if a
line was exactly 8190 plus a multiple of 8192 characters long the config
parser would skip the following line.
Additionally fix comment in voicemail.conf sample config. It previously
stated that emailbody can only contain up to 512 characters which is
always wrong. The buffer is normally 8192 characters unless LOW_MEMORY
is enabled then it is 512 characters. The updated comment states that
the line can be up to 8190 or 510 characters since the line feed and
NULL terminator each use a character.
Richard Mudgett [Thu, 22 Feb 2018 19:53:20 +0000 (13:53 -0600)]
res_pjsip_refer.c: Fix attended transfer race condition crash.
The transferrer's session channel was destroyed by the transferrer's
serializer thread in a race condition with the transfer target's
serializer thread during an attended transfer. The transfer target's
serializer was attempting to clean up a deferred end status on behalf of
the transferrer's channel when it should have passed the action to the
transferrer's serializer. When the transfer target's serializer lost the
race then both threads wind up trying to end the transferrer's session.
* Push the ast_sip_session_end_if_deferred() call onto the transferrer's
serializer to avoid a race condition that results in a crash. The
session_end() function that could be called by
ast_sip_session_end_if_deferred() really must be executed by the
transferrer's serializer to avoid this kind of crash.
Kevin Harwell [Thu, 22 Feb 2018 17:15:55 +0000 (11:15 -0600)]
AMI: Bumping AMI non-breaking number for Asterisk 15.3.0 release
A few changes were made to AMI:
* Fixed "(null):" header in AMI AsyncAGIEnd event
* A mute header was added to the ConfbridgeJoin AMI event
* ConfbridgeList action's ConfbridgeList events now output all
the standard channel snapshot headers
Sean Bright [Mon, 5 Feb 2018 22:46:26 +0000 (17:46 -0500)]
AST-2018-006: Properly handle WebSocket frames with 0 length payload.
In ast_websocket_read() we were not adequately checking that the
payload_len was non-zero before passing it to ws_safe_read(). Calling
ws_safe_read with a len argument of 0 will result in a busy loop until
the underlying socket is closed.
Kevin Harwell [Wed, 31 Jan 2018 19:37:54 +0000 (13:37 -0600)]
AST-2018-003: Crash with an invalid SDP fmtp attribute
pjproject's fmtp retrieval function failed to catch invalid fmtp attributes.
Because of this Asterisk would crash if given an SDP with an invalid fmtp
attribute.
When retrieving the format this patch now makes sure the fmtp attribute is
available. If not available it now returns an error status.
Kevin Harwell [Wed, 31 Jan 2018 19:33:16 +0000 (13:33 -0600)]
AST-2018-002: Crash with an invalid SDP media format description
pjproject's media format parsing algorithm failed to catch invalid values.
Because of this Asterisk would crash if given an SDP with a invalid media
format description.
When parsing the media format description this patch now properly parses the
value and returns an error status if it can't successfully parse/convert the
value.
George Joseph [Tue, 6 Feb 2018 18:07:18 +0000 (11:07 -0700)]
AST-2018-005: res_pjsip_transport_management: Move to core
Since res_pjsip_transport_management provides several attack
mitigation features, its functionality moved to res_pjsip and
this module has been removed. This way the features will always
be available if res_pjsip is loaded.
George Joseph [Tue, 6 Feb 2018 17:28:49 +0000 (10:28 -0700)]
AST-2018-005: Fix tdata leaks when calling pjsip_endpt_send_response(2)
pjsip_distributor:
authenticate() creates a tdata and uses it to send a challenge or
failure response. When pjsip_endpt_send_response2() succeeds, it
automatically decrements the tdata ref count but when it fails, it
doesn't. Since we weren't checking for a return status, we weren't
decrementing the count ourselves on error and were therefore leaking
tdatas.
res_pjsip_session:
session_reinvite_on_rx_request wasn't decrementing the ref count
if an error happened while sending a 491 response.
pre_session_setup wasn't decrementing the ref count if
while sending an error after a pjsip_inv_verify_request failure.
res_pjsip:
ast_sip_send_response wasn't decrementing the ref count on error.
George Joseph [Tue, 6 Feb 2018 17:21:32 +0000 (10:21 -0700)]
AST-2018-005: Add a check for NULL tdata in ast_sip_failover_request
It was discovered that there are some corner cases where a pjsip tsx
might have no last_tx so calling ast_sip_failover_request with
a NULL last_tx as its tdata would cause a crash.
Joshua Colp [Wed, 7 Feb 2018 14:09:14 +0000 (14:09 +0000)]
AST-2018-004: Restrict the number of Accept headers in a SUBSCRIBE.
When receiving a SUBSCRIBE request the Accept headers from it are
stored locally. This operation has a fixed limit of 32 Accept headers
but this limit was not enforced. As a result it was possible for
memory outside of the allocated space to get written to resulting
in a crash.
This change enforces the limit so only 32 Accept headers are
processed.
Joshua Colp [Sat, 13 Jan 2018 14:04:33 +0000 (14:04 +0000)]
AST-2018-001: rtp / channel: Don't allow an unnegotiated format to be passed up.
When an RTP packet is received by an RTP engine it has to map the
payload into the Asterisk format. The code was incorrectly checking
our own static list for ALL payloads if it couldn't find a negotiated one.
This included dynamic payloads. If the payload mapped to a format
of a different type (for example receiving a video packet on an audio
RTP instance) then the core stream code could cause a crash if a legacy
channel driver was in use as no stream would be present.
To provide further protection the core stream code will no longer assume
that a video or audio frame will always have a stream for legacy channel
drivers. If no stream is present the frame is dropped.
Joshua Colp [Tue, 20 Feb 2018 16:33:43 +0000 (12:33 -0400)]
chan_sip: Emit a second ringing event to ensure channel is found.
When constructing a dialog-info+xml NOTIFY message a ringing channel
is found if the state is ringing and further information is placed into
the message. Due to the migration to the Stasis message bus this did
not always work as expected.
This change raises a second ringing event in such a way to guarantee
that the event is received by chan_sip and another lookup is done to
find the ringing channel.
projects / thirdparty / asterisk.git / log
