Richard Mudgett [Fri, 4 Dec 2015 21:36:45 +0000 (15:36 -0600)]
chan_sip: Fix crash involving the bogus peer during sip reload.
A crash happens sometimes when performing a CLI "sip reload". The bogus
peer gets refreshed while it is in use by a new call which can cause the
crash.
* Protected the global bogus peer object with an ao2 global object
container.
Joshua Colp [Fri, 4 Dec 2015 14:15:24 +0000 (10:15 -0400)]
Fix crash in audiohook translate to slin
This patch fixes a crash which would occur when an audiohook was
applied to a channel using an audio codec that could not be translated
to signed linear (such as when using pass-through codecs like OPUS or
when the codec translator module for the format in use is not loaded).
Richard Mudgett [Mon, 30 Nov 2015 22:42:47 +0000 (16:42 -0600)]
sched.c: Make not return a sched id of 0.
According to the API doxygen a sched ID of 0 is valid. Unfortunately, 0
was never returned historically and several users incorrectly coded usage
of the returned sched ID assuming that 0 was invalid.
Richard Mudgett [Tue, 24 Nov 2015 18:44:53 +0000 (12:44 -0600)]
Audit improper usage of scheduler exposed by 5c713fdf18f.
channels/chan_iax2.c:
* Initialize struct chan_iax2_pvt scheduler ids earlier because of
iax2_destroy_helper().
channels/chan_sip.c:
channels/sip/config_parser.c:
* Fix initialization of scheduler id struct members. Some off nominal
paths had 0 as a scheduler id to be destroyed when it was never started.
chan_skinny.c:
* Fix some scheduler id comparisons that excluded the valid 0 id.
channel.c:
* Fix channel initialization of the video stream scheduler id.
pbx_dundi.c:
* Fix channel initialization of the packet retransmission scheduler id.
Kevin Harwell [Wed, 25 Nov 2015 21:26:35 +0000 (15:26 -0600)]
fastagi: record file closed after sending result
The fastagi record-file testsuite test sometimes fails reporting an empty
recorded file. This was happening because Asterisk was sending the agi result
notification prior to actually closing the file and the data, being buffered,
had not been written to the file yet when the test attempts to check the file
size.
This patch makes it so the record file stream is closed prior to sending the
agi result notification.
Walter Doekes [Wed, 25 Nov 2015 19:29:30 +0000 (20:29 +0100)]
main: Slight refactor of main. Improve color situation.
Several issues are addressed here:
- main() is large, and half of it is only used if we're not rasterisk;
fixed by spliting up the daemon part into a separate function.
- Call ast_term_init from rasterisk as well.
- Remove duplicate code reading/writing asterisk history file.
- Attempt to tackle background color issues and color changes that
occur. Tested by starting asterisk -c until the colors stopped
changing at odd locations.
Joshua Colp [Sat, 14 Nov 2015 13:02:10 +0000 (09:02 -0400)]
hashtab: Add NULL check when destroying iterator.
The hashtab API is pretty NULL tolerant which has resulted
in remaining callers not doing much checks themselves.
Unfortunately the function to destroy an iterator does not
do a NULL check and will result in a crash if passed NULL.
This change fixes that.
Steve Davies [Wed, 11 Nov 2015 10:16:22 +0000 (10:16 +0000)]
Further fixes to improper usage of scheduler
When ASTERISK-25449 was closed, a number of scheduler issues mentioned in
the comments were missed. These have since beed raised in ASTERISK-25476
and elsewhere.
This patch attempts to collect all of the scheduler issues discovered so
far and address them sensibly.
Alexander Traud [Tue, 10 Nov 2015 15:29:35 +0000 (16:29 +0100)]
rtp_engine: Init a format-attribute module to its RFC defaults.
Previously, format-attribute modules relied on an existing fmtp line in SDP
negotiation. However, fmtp is optional for several formats like the Opus Codec.
Now, the format-attribute module is called with an empty fmtp, which allows the
module to initialise itself to RFC defaults. Furthermore now, Asterisk is able
to differentiate between internally and externally created formats.
Walter Doekes [Fri, 6 Nov 2015 13:54:59 +0000 (14:54 +0100)]
func_callerid: Document that CALLERID(pres) is available.
CALLERPRES() says that it's deprecated in favor of CALLERID(num-pres)
and CALLERID(name-pres). But for channel driver that don't make a
distinction between the two (e.g. SIP), it makes more sense to get/set
both at once. This change reveals the availability of CALLERID(pres),
CONNECTEDLINE(pres), REDIRECTING(orig-pres), REDIRECTING(to-pres) and
REDIRECTING(from-pres).
Walter Doekes [Fri, 6 Nov 2015 13:36:40 +0000 (14:36 +0100)]
xmldoc: Improve xmldoc wrapping of 'core show ...' output.
Previously, the wrapping did both lookahead and lookback, which,
together with color escape sequences, caused some lines to be wrapped
way earlier than other lines. This led to inconsistent output.
This simplifies the wrapping code and makes it more sane: if maxcolumns
is hit, we simply jump back to the last space and wrap there.
Corey Farrell [Wed, 4 Nov 2015 14:25:52 +0000 (09:25 -0500)]
Fix cli display of build options.
A previous commit reduced the AST_BUILDOPTS compiler define to
only include options that affected ABI. This included some options
that were previously displayed by cli "core show settings". This
change corrects the CLI display while still restricting buildopts.h
to ABI effecting options only.
Alexander Traud [Wed, 21 Oct 2015 13:06:29 +0000 (15:06 +0200)]
format: Update the maximum packetization time for iLBC 30.
In September 2006, the maximum packetization time (ptime) were set to such a
low value, packetization was disabled for many codecs actually. This was fixed
for many codecs but not for iLBC 30. This enables packetization for iLBC which
can be enabled for example via allow=ilbc:60,gsm,alaw,ulaw in the file sip.conf.
On v13, loading several thousand PJSIP endpoints on Asterisk start causes
a deadlock most of the time.
Thanks to mdu113 for discovering that there was a call to pgsql_exec() not
protected by the pgsql_lock reentrancy lock.
{quote}
I believe a code path exists that attempts to use pgsql connection without
locking pgsql_lock. I believe what happens during that deadlock that I
see is two concurrent threads are both attempting to send query to pgsql,
one of the thread is using a code path without locking pgsql_lock. If
they managed to send queries at the same time, it seems postgres ignores
one of the queries and replies only to the one of them. If it happens so
that the thread holding the lock didn't receive the reply it will wait for
it (and hold the lock) forever (or at least for very long time), thus
completely blocking all access to db.
{quote}
* Added missing reentrancy locking around pgsql_exec() in find_table().
* Moved unlock of pgsql_lock in unload_module() to avoid locking inversion
between the psql_tables list lock and the pgsql_lock.
Richard Mudgett [Mon, 12 Oct 2015 16:20:29 +0000 (11:20 -0500)]
config.c: Fix potential memory corruption after [section](+).
The memory corruption could happen if the [section](+) is the last section
in the file with trailing comments. In this case process_text_line() has
left *last_cat is set to newcat and newcat is destroyed.
Richard Mudgett [Mon, 12 Oct 2015 16:21:55 +0000 (11:21 -0500)]
config.c: Fix #include after [section](+).
An #include right after a [section](+) would associate any variable
assignments before a new section in the #include with the wrong section.
* Fix section association by setting the current section to the appended
section.
* Fix '+' and '!' section flag interaction corner case depending upon
which flag came first. If the '!' came first then it would be ignored.
If the '!' came after then it would affect the appended section. The '!'
will now no longer be ignored.
Matt Jordan [Wed, 7 Oct 2015 01:43:58 +0000 (20:43 -0500)]
res/res_rtp_asterisk: Fix assignment after ao2 decrement
When we decide we will no longer schedule an RTCP write, we remove the
reference to the RTP instance, then assign -1 to the stored scheduler ID
in case something else comes along and wants to see if anything is scheduled.
That scheduler ID is on the RTP instance. After 60a9172d7ef2 was merged to
fix the regression introduced by 3cf0f29310, this improper assignment on a
potentially destroyed object started getting tripped on the build agents.
Frankly, this should have been crashing a lot more often earlier. I can only
assume that the timing was changed just enough by both changes to start
actually hitting this problem.
As it is, simply moving the assignment prior to the ao2 deference is sufficient
to keep the RTP instance from being referenced when it is very, truly,
aboslutely dead.
(Note that it is still good practice to assign -1 to the scheduler ID when we
know we won't be scheduling it again, as the ao2 deref *may* not always destroy
the ao2 object.)
chan_sip: Fix port parsing for IPv6 addresses in SIP Via headers.
If a Via header containes an IPv6 address and a port number is ommitted,
as it is the standard port, we now leave the port empty and to not set it
to the value after the first colon of the IPv6 address.
Matt Jordan [Tue, 6 Oct 2015 02:34:41 +0000 (21:34 -0500)]
Fix improper usage of scheduler exposed by 5c713fdf18f
When 5c713fdf18f was merged, it allowed for scheduled items to have an ID of
'0' returned. While this was valid per the documentation for the API, it was
apparently never returned previously. As a result, several users of the
scheduler API viewed the result as being invalid, causing them to reschedule
already scheduled items or otherwise fail in interesting ways.
This patch corrects the users such that they view '0' as valid, and a returned
ID of -1 as being invalid.
Note that the failing HEP RTCP tests now pass with this patch. These tests
failed due to a duplicate scheduling of the RTCP transmissions.
Ivan Poddubny [Sat, 3 Oct 2015 11:27:27 +0000 (14:27 +0300)]
manager: Fix GetConfigJSON returning invalid JSON
When GetConfigJSON was introduced back in 1.6, it returned each
section as an array of strings: ["key=value", "key2=value2"].
Afterwards, it was changed a few times and became
["key": "value", "key2": "value2"], which is not a correct JSON.
This patch fixes that by constructing a JSON object {} instead of
an array [].
Ivan Poddubny [Mon, 28 Sep 2015 06:36:30 +0000 (09:36 +0300)]
channel.c: Fix NewCallerid AMI event not been sent on Caller ID change
Currently, NewCallerid is sent only when pointers to number or name
strings change, which is not always the case. The newly allocated string
may use the same memory, so pointers match, while the content
is different. As a result, Caller ID updates are often not reported.
With this patch, actual strings are compared, not the pointers.
core/logging: Fix logging to more than one syslog channel
Currently, Asterisk will log to the last configured syslog
channel in logger.conf. This is due to the fact that the
final call to openlog() supersedes all of the previous calls.
This commit removes the call to openlog() and passes the
facility to ast_log_vsyslog(), along with utilizing the
LOG_MAKEPRI macro to ensure that the message is routed to
the correct facility and with the correct priority.
pbx: Update device and presence state when changing a hint extension.
When changing a hint extension without removing the hint first the
device state and presence state is not updated. This causes the state
of the hint to be that of the previous extension and not the current
one. This state is kept until a state change occurs as a result of
something (presence state change, device state change).
This change updates the hint with the current device and presence
state of the new extension when it is changed. Any state callbacks
which may have been added before the hint extension is changed are
also informed of the new device and presence state if either have
changed.
Walter Doekes [Thu, 17 Sep 2015 09:52:09 +0000 (11:52 +0200)]
chan_sip: Fix From header truncation for extremely long CALLERID(name).
The CALLERID(num) and CALLERID(name) and other info are placed into the
`char from[256]` in initreqprep. If the name was too long, the addr-spec
and params wouldn't fit.
Code is moved around so the addr-spec with params is placed there first,
and then fitting in as much of the display-name as possible.
Alexander Traud [Fri, 28 Aug 2015 21:06:02 +0000 (23:06 +0200)]
translate: Fix transcoding while different in frame size.
When Asterisk translates between codecs each with a different frame size (for
example between iLBC 30 and Speex-WB), too large frames were created by
ast_trans_frameout. Now, ast_trans_frameout is called with the correct frame
length, creating several frames when necessary. Affects all transcoding modules
which used ast_trans_frameout: GSM, iLBC, LPC10, and Speex.
Mark Michelson [Thu, 10 Sep 2015 22:19:26 +0000 (17:19 -0500)]
scheduler: Use queue for allocating sched IDs.
It has been observed that on long-running busy systems, a scheduler
context can eventually hit INT_MAX for its assigned IDs and end up
overflowing into a very low negative number. When this occurs, this can
result in odd behaviors, because a negative return is interpreted by
callers as being a failure. However, the item actually was successfully
scheduled. The result may be that a freed item remains in the scheduler,
resulting in a crash at some point in the future.
The scheduler can overflow because every time that an item is added to
the scheduler, a counter is bumped and that counter's current value is
assigned as the new item's ID.
This patch introduces a new method for assigning scheduler IDs. Instead
of assigning from a counter, a queue of available IDs is maintained.
When assigning a new ID, an ID is pulled from the queue. When a
scheduler item is released, its ID is pushed back onto the queue. This
way, IDs may be reused when they become available, and the growth of ID
numbers is directly related to concurrent activity within a scheduler
context rather than the uptime of the system.
David M. Lee [Fri, 4 Sep 2015 21:33:39 +0000 (16:33 -0500)]
res_rtp_asterisk: Add more ICE debugging
In working through a recent ICE negotiation bug, I found the debug
logging in res_rtp_asterisk to be lacking. This patch adds a number of
debug and warning statements that were helpful.
David M. Lee [Fri, 4 Sep 2015 21:06:39 +0000 (16:06 -0500)]
Fix when remote candidates exceed PJ_ICE_MAX_CAND
We were passing the wrong count into pj_ice_sess_create_check_list(),
causing the create to fail if we ever received more than PJ_ICE_MAX_CAND
candidates.
Joshua Colp [Wed, 26 Aug 2015 10:40:32 +0000 (07:40 -0300)]
chan_sip: Allow call pickup to set the hangup cause.
The call pickup implementation in chan_sip currently sets the channel
hangup cause to "normal clearing" if call pickup is successfully
performed. This action overwrites the "answered elsewhere" hangup cause
set by the call pickup code and can result in the SIP device in
question showing a missed call when it should not.
This change sets the hangup cause to "normal clearing" as a
default initially but allows the call pickup to change it as
needed.
contrib: script install_prereq should install sqlite3
Asterisk needs the sqlite 3 library, which is package
sqlite-devel in CentOS. By adding this package to the
script, a problem with configure failing is resolved.
Richard Mudgett [Mon, 17 Aug 2015 21:41:19 +0000 (16:41 -0500)]
app_queue.c: Extract some functions for simpler code.
* Extract set_queue_member_pause() from set_member_paused() for simpler
and more consistent code.
* Extract set_queue_member_ringinuse() from
set_member_ringinuse_help_members() for simpler code.
NOTE: This may fix a consistency issue with realtime ringinuse
because the ordering of things was backported from v13. It is
similar to how set_member_paused() treats realtime for paused.
Richard Mudgett [Fri, 14 Aug 2015 17:55:28 +0000 (12:55 -0500)]
app_queue.c: Fix setting QUEUE_MEMBER 'paused' and 'ringinuse'.
Setting the 'paused' and 'ringinuse' options on a queue member using the
dialplan function QUEUE_MEMBER did not behave the same way as the
equivalent dialplan applications or AMI actions.
* Made queue_function_mem_write() call the set_member_paused() and
set_member_value() for the 'paused' and 'ringinuse' options respectively.
A beneficial side effect is that the queue name is now optional and sets
the value in all queues the interface is a member.
NOTE: This may fix a consistency issue with the realtime paused setting
since how the value is set is controlled by set_member_paused() which
treats realtime a little better.
Kevin Harwell [Wed, 12 Aug 2015 17:59:53 +0000 (12:59 -0500)]
chan_sip.c: wrong peer searched in sip_report_security_event
In chan_sip, after handling an incoming invite a security event is raised
describing authorization (success, failure, etc...). However, it was doing
a lookup of the peer by extension. This is fine for register messages, but
in the case of an invite it may search and find the wrong peer, or a non
existent one (for instance, in the case of call pickup). Also, if the peers
are configured through realtime this may cause an unnecessary database lookup
when caching is enabled.
This patch makes it so that sip_report_security_event searches by IP address
when looking for a peer instead of by extension after an invite is processed.
Richard Mudgett [Mon, 10 Aug 2015 18:43:19 +0000 (13:43 -0500)]
chan_dahdi.c: Flush the DAHDI write buffer after starting DTMF.
Pressing DTMF digits on a phone to go out on a DAHDI channel can result in
the digit not being recognized or even heard by the peer.
Phone -> Asterisk -> DAHDI/channel
Turns out the DAHDI behavior with DTMF generation (and any other generated
tones) is exposed by the "buffers=" setting in chan_dahdi.conf. When
Asterisk requests to start sending DTMF then DAHDI waits until its write
buffer is empty before generating any samples for the DTMF tones. When
Asterisk subsequently requests DAHDI to stop sending DTMF then DAHDI
immediately stops generating the DTMF samples. As a result, the more
samples there are in the DAHDI write buffer the shorter the time DTMF
actually gets sent on the wire. If there are more samples in the write
buffer than the time DTMF is supposed to be sent then no DTMF gets sent on
the wire. With the "buffers=12,half" setting and each buffer representing
20 ms of samples then the DAHDI write buffer is going to contain around
120 ms of samples. For DTMF to be recognized by the peer the actual sent
DTMF duration needs to be a minimum of 40 ms. Therefore, the intended
duration needs to be a minimum of 160 ms for the peer to receive the
minimum DTMF digit duration to recognize it.
A simple and effective solution to work around the DAHDI behavior is for
Asterisk to flush the DAHDI write buffer when sending DTMF so the full
duration of DTMF is actually sent on the wire. When someone is going to
send DTMF they are not likely to be talking before sending the tones so
the flushed write samples are expected to just contain silence.
* Made dahdi_digit_begin() flush the DAHDI write buffer after requesting
to send a DTMF digit.
Richard Mudgett [Wed, 5 Aug 2015 19:21:50 +0000 (14:21 -0500)]
chan_dahdi.c: Lock private struct for ast_write().
There is a window of opportunity for DTMF to not go out if an audio frame
is in the process of being written to DAHDI while another thread starts
sending DTMF. The thread sending the audio frame could be past the
currently dialing check before being preempted by another thread starting
a DTMF generation request. When the thread sending the audio frame
resumes it will then cause DAHDI to stop the DTMF tone generation. The
result is no DTMF goes out.
* Made dahdi_write() lock the private struct before writing to the DAHDI
file descriptor.
Joshua Colp [Tue, 11 Aug 2015 10:24:30 +0000 (07:24 -0300)]
res_http_websocket: Forcefully terminate on write errors.
The res_http_websocket module will currently attempt to close
the WebSocket connection if fatal cases occur, such as when
attempting to write out data and being unable to. When the
fatal cases occur the code attempts to write a WebSocket close
frame out to have the remote side close the connection. If
writing this fails then the connection is not terminated.
This change forcefully terminates the connection if the
WebSocket is to be closed but is unable to send the close frame.
Joshua Colp [Wed, 5 Aug 2015 10:23:21 +0000 (07:23 -0300)]
res_rtp_asterisk: Don't leak temporary key when enabling PFS.
A change recently went in which enabled perfect forward secrecy for
DTLS in res_rtp_asterisk. This was accomplished two different ways
depending on the availability of a feature in OpenSSL. The fallback
method created a temporary instance of a key but did not free it.
This change fixes that.
Mark Duncan [Tue, 28 Jul 2015 10:33:39 +0000 (19:33 +0900)]
res/res_rtp_asterisk: Add ECDH support
This will add ECDH support to Asterisk. It will
detect auto ECDH support in OpenSSL
(1.0.2b and above) during ./configure. If this is
available, it will use it,
otherwise it will fall back to prime256v1 (this
behavior is consistent with
other projects such as Apache and nginx).
This fixes WebRTC being broken in Firefox 38+ due
to Firefox now only supporting
ciphers with perfect forward secrecy.
Mark Michelson [Wed, 29 Jul 2015 19:35:58 +0000 (14:35 -0500)]
res_http_websocket: Properly encode 64 bit payload
A test agent was continuously failing all ARI tests when run against
Asterisk 13. As it turns out, the reason for this is that on those test
runs, for some reason we decided to use the super extended 64 bit
payload length for websocket text frames instead of the extended 16 bit
payload length. For 64-bit payloads, the expected byte order over the
network is
7, 6, 5, 4, 3, 2, 1, 0
However, we were sending the payload as
3, 2, 1, 0, 7, 6, 5, 4
This meant that we were saying to expect an absolutely MASSIVE payload
to arrive. Since we did not follow through on this expected payload
size, the client would sit patiently waiting for the rest of the payload
to arrive until the test would time out.
With this change, we use the htobe64() function instead of htonl() so
that a 64-bit byte-swap is performed instead of a 32 bit byte-swap.
Mark Michelson [Thu, 23 Jul 2015 18:11:42 +0000 (13:11 -0500)]
Local channels: Alternate solution to ringback problem.
Commit 54b25c80c8387aea9eb20f9f4f077486cbdf3e5d solved an issue where a
specific scenario involving local channels and a native local RTP bridge
could result in ringback still being heard on a calling channel even
after the call is bridged.
That commit caused many tests in the testsuite to fail with alarming
consequences, such as not sending DialBegin and DialEnd events, and
giving incorrect hangup causes during calls.
This commit reverts the previous commit and implements and alternate
solution. This new solution involves only passing AST_CONTROL_RINGING
frames across local channels if the local channel is in AST_STATE_RING.
Otherwise, the frame does not traverse the local channels. By doing
this, we can ensure that a playtones generator does not get started on
the calling channel but rather is started on the local channel on which
the ringing frame was initially indicated.
Mark Michelson [Wed, 22 Jul 2015 14:46:29 +0000 (09:46 -0500)]
Local channels: Do not block control -1 payloads.
Control frames with a -1 payload are used as a special signal to stop
playtones generators on channels. This indication is sent both by
app_dial as well as by ast_answer() when a call is answered in case any
tones were being generated on a calling channel.
This control frame type was made to stop traversing local channel pairs
as an optimization, because it was thought that it was unnecessary to
send these indications, and allowing such unnecessary control frames to
traverse the local channels would cause the local channels to optimize
away less quickly.
As it turns out, through some special magic dialplan code, it is
possible to have a tones being played on a non-local channel, and it is
important for the local channel to convey that the tones should be
stopped. The result of having tones continue to be played on the
non-local channel is that the tones play even once the channel has been
bridged. By not blocking the -1 control frame type, we can ensure that
this situation does not happen.
audiohook: Read the correct number of samples based on audiohook format.
Due to changes in audiohooks to support different sample rates the
underlying storage of samples is in the format of the audiohook
itself and not of the format being requested. This means that if a
channel is using G722 the samples stored will be at 16kHz. If
something subsequently reads from the audiohook at a format which
is not the same sample rate as the audiohook the number of samples
needs to be adjusted.
Given the following example:
1. Channel writing into audiohook at 16kHz (as it is using G722).
2. Chanspy reading from audiohook at 8kHz.
The original code would read 160 samples from the audiohook for
each 20ms of audio. This is incorrect. Since the audio in the
audiohook is at 16kHz the actual number needing to be read is 320.
Failure to read this much would cause the audiohook to reset
itself constantly as the buffer became full.
This change adjusts the requested number of samples by determining
the duration of audio requested and then calculating how many
samples that would be in the audiohook format.
Rusty Newton [Mon, 20 Jul 2015 22:16:36 +0000 (17:16 -0500)]
Documentation: chan_sip doesn't support WS or WSS in outbound register.
* In sip.conf.sample fix sentence where we said that WS or WSS are supported
transports for use in an outbound register definition. They are not
supported in that case.
Matt Jordan [Sat, 27 Jun 2015 22:53:37 +0000 (17:53 -0500)]
tests/test_devicestate: Add additional tests for the device state API
This patch adds more tests that exercise the device state API. This includes:
* Tests that cover adding a device state provider, as well as deleting a
device state provider. This also verifies that you cannot add an
already added device state provider, and cannot delete an already
deleted device state provider.
* A test that covers changing device state and receiving said updates
from a device state subscriber. This also covers hitting both the
device state cache as well as a custom device state provider.
* A test that covers converting device state to channel state and device
state values to a string representation and back.
* A test that covers obtaining device state from an active channel and a
channel driver that provides its own device state.
Matt Jordan [Sat, 27 Jun 2015 22:51:43 +0000 (17:51 -0500)]
main/devicestate: Prevent duplicate registration of device state providers
Currently, the device state provider API will allow you to register a
device state provider with the same case insensitive name more than
once. This could cause strange issues, as the duplicate device state
providers will not be queried when a device's state has to be polled.
This patch updates the API such that a device state provider with the
same name as one that has already registered will be rejected.
res_rtp_asterisk: Prevent simultaneous access to DTLS SSL context.
This change moves logic for setting up the DTLS SSL contexts to
when the SDP is done being processed instead of when ICE negotiation
completes. It also stops handshakes from being initiated when we
are acting as a server.
Manipulating the SSL context when ICE negotiation has completed
is problematic as the SSL context is not protected and if acting
as a client the remote side may have started DTLS negotiation
already.
The retransmission timeout timer code has also been split up
and simplified some. Both RTP and RTCP now have their own timers
and the points at which the timer is stopped and started is now
more specific. When a packet is sent the timer is started. When
a response is received but before it is processed the timer is
stopped. This provides a guarantee that the timeout is not
occurring while the response is processed.
Walter Doekes [Thu, 2 Jul 2015 11:25:36 +0000 (13:25 +0200)]
chan_sip: Fix early call pickup caused deadlock.
If non-magic pickup (no "pickup-" in callid) is used, chan_sip locks the
channel it wants to pick up, and a bit further down, it locks the
channel list when allocating a new channel.
That causes a deadlock when another part of the code traverses over the
channel list, locking all the channels one by one.
This changeset fixes it by releasing the locks before calling sip_new
and reacquiring them afterwards. Unfortunately this involves doing the
checks we already did again (because the channel may have changed).
While trying to avoid duplicate code, I did some refactoring for
readability:
- if refer_locked == 1, we guarantee there is a locked channel
- magic_callid holds a cached version of !ast_strlen_zero(pickup.exten)
This is for branch 11 only. It appears that the changed code in 13 does
not lock the components like it does in 11 and below. Reproducing the
deadlock on 13 has thusfar failed.
projects / thirdparty / asterisk.git / log
