Kevin Harwell [Thu, 13 Feb 2020 21:08:10 +0000 (15:08 -0600)]
res_rtp_asterisk: bad audio (static) due to incomplete dtls/srtp setup
There was a race condition between client initiated DTLS setup, and handling
of server side ice completion that caused the underlying SSL object to get
cleared during DTLS initialization. If this happened Asterisk would be left
in a partial DTLS setup state. RTP packets were sent and received, but were
not being encrypted and decrypted. This resulted in no audio, or static.
Specifically, this occurred when '__rtp_recvfrom' was processing the handshake
sequence from the client to the server, and then 'ast_rtp_on_ice_complete'
gets called from another thread and clears the SSL object when calling the
'dtls_perform_setup' function. The timing had to be just right in the sense
that from the external SSL library perspective SSL initialization completed
(rtp recv), Asterisk clears/resets the SSL object (ice done), and then checks
to see if SSL is intialized (rtp recv). Since it was cleared, Asterisk thinks
it is not finished, thus not completing 'dtls_srtp_setup'.
This patch removes calls to 'dtls_perform_setup', which clears the SSL object,
in 'ast_rtp_on_ice_complete'. When ice completes, there is no reason to clear
the underlying SSL object. If an ice candidate changes a full protocol level
renegotiation occurs. Also, in the case of bundled ICE candidates are reused
when a stream is added. So no real reason to have to clear, and reset in this
instance.
Also, this patch adds a bit of extra logging to aid in diagnosis of any future
problems.
Sean Bright [Tue, 11 Feb 2020 13:46:02 +0000 (08:46 -0500)]
res_musiconhold: Avoid spurious warning when 'format' is the empty string
The change to res_config_odbc that allowed empty strings to be
returned to realtime consumers¹ causes a warning to be emitted when
loading MoH classes. So we need to treat an empty 'format' as if it
was not specified to avoid the warning.
Given a scenario where session refreshes occur close to
each other while another is finishing it was possible for
the session refreshes to occur out of order. It was
also possible for session refreshes to be delayed for
quite some time if a session refresh did not result in
a topology change.
For the out of order session refreshes the first session
refresh would be queued due to a transaction in progress.
This transaction would then finish. When finished a
separate task to process the delayed requests queue
would be queued for handling. A second refresh would
be requested internally before this delayed request
queued task was processed. As no transaction was in
progress this session refresh would be immediately
handled before the queued session refresh.
The code will now check if any delayed requests exist
before allowing a session refresh to immediately occur.
If any exist then the session refresh is queued.
For the delayed session refreshes if a session refresh
did not result in a topology change the attempt would
be immediately stopped and no other delayed requests would
be processed.
The code will now go through the entire delayed requests
queue until a delayed request results in a request
actually being sent.
Currently aptitude is installed using interactive mode. This patch
changes this to use the non-interactive mode as it can block
automatic dependencies installation, ex: CI, Docker build.
Walter Doekes [Tue, 28 Jan 2020 08:34:45 +0000 (09:34 +0100)]
chan_sip: Clarify in sample docs how directmediapermit/-acl should be used
It said "restrict [...] which peers should be able to pass [audio]
to each other".
However, these settings are not global (for which you would expect
signaling IPs to be checked). These settings are available per peer
only, and the IPs being checked, are the RTP IPs.
Joshua C. Colp [Tue, 4 Feb 2020 14:18:13 +0000 (10:18 -0400)]
res_rtp_asterisk: Don't produce transport-cc if no packets.
The code assumed that when the transport-cc feedback
function was called at least one packet will have been
received. In practice this isn't always true, so now
we just reschedule the sending and do nothing.
George Joseph [Mon, 3 Feb 2020 16:24:58 +0000 (09:24 -0700)]
message.c: Add option to suppress the Message channel AMI and ARI events
In order to reduce the amount of AMI and ARI events generated,
the global "Message/ast_msg_queue" channel can be set to suppress
it's normal channel housekeeping events such as "Newexten",
"VarSet", etc. This can greatly reduce load on the manager
and ARI applications when the Digium Phone Module for Asterisk
is in use. To enable, set "hide_messaging_ami_events" in
asterisk.conf to "yes" In Asterisk versions <18, the default
is "no" preserving existing behavior. Beginning with
Asterisk 18, the option will default to "yes".
NOTE: This change does not affect UserEvents or the ARI
TextMessageReceived events.
* Added the "hide_messaging_ami_events" option to asterisk.conf.
* Changed message.c to set the AST_CHAN_TP_INTERNAL property on
the "Message/ast_msg_queue" channel if the option is set in
asterisk.conf. This suppresses the reporting of the events.
Sean Bright [Wed, 29 Jan 2020 14:57:38 +0000 (09:57 -0500)]
res_config_odbc: Preserve empty strings returned by the database
When res_config_odbc (and perhaps other realtime backends) reads a SQL
NULL from the database, it coalesces the value to the empty string
which prevents it from being returned to the realtime core.
However, if it instead reads the empty string from the database, it
needs a way to encode that fact without having the value omitted
entirely. It does this by changing the value to a string with a single
space. The realtime code in main/config.c recognizes this special case
and _turns the string back into the empty string_ before passing it to
realtime API consumers.
For all of this to work, we need to ensure that we actually pass the
single-space-string back to the realtime core, which is currently
failing because we are trimming the value before checking its
content. So instead we now special case the single-space-string case
so that empty values are returned properly.
Sean Bright [Tue, 28 Jan 2020 19:23:19 +0000 (14:23 -0500)]
res_stasis_playback: Prevent media_index from going out of bounds
Incrementing stasis_app_playback.media_index directly in our playback
loop means that when we reach the end of our playlist the index into
the vector will be outside of the bounds of the vector.
Instead use a temporary variable and only assign when we're sure that
we are in bounds.
Joshua C. Colp [Tue, 28 Jan 2020 15:18:45 +0000 (15:18 +0000)]
res_pjsip_pubsub: Increment persistence data ref when recreating.
Each subscription needs to have a reference to the persisted data
for it, as well as the main JSON contained within the tree. When
recreating a subscription this did not occur and they both shared
the same reference.
Kevin Harwell [Mon, 27 Jan 2020 18:01:15 +0000 (12:01 -0600)]
stasis/app: don't lock an app before a call to send
Calling 'app_send' eventually calls the app's message handler. It's possible
for a handler to obtain a lock on another object, and then need/want to lock
the app object. If the caller of 'app_send' locks the app object prior to
calling then there's a potential for a deadlock, if another thread calls
'app_send' without locking.
This patch makes it so 'app_send' is not called with the app object locked in
the section of code doing such.
Kevin Harwell [Mon, 27 Jan 2020 17:44:45 +0000 (11:44 -0600)]
res_stasis: trigger cleanup after update
The cleanup code in stasis shuts down applications if they are in a deactivated
state, and no longer have explicit subscriptions. When registering an app the
cleanup code was running before calling 'update'. When it should be executed
after 'update' since a call to register may re-activate the app. We don't want
it to shutdown before the 'update' otherwise the app won't be re-activated,
or registered.
This patch makes it so the cleanup code is executed post 'update'.
Walter Doekes [Mon, 13 Jan 2020 10:13:41 +0000 (11:13 +0100)]
chan_sip: Always process updated SDP on media source change
Fixes no-audio issues when the media source is changed and
strictrtp is enabled (default).
If the peer media source changes, the SDP session version also changes.
If it is lower than the one we had stored, chan_sip would ignore it.
This changeset keeps track of the remote media origin identifier,
comparing that as well. If it changes, the session version needn't be
higher for us to accept the SDP.
Common scenario where this would've caused problems: a separate media
gateway that informs the caller about premium rates before handing off
the call to the final destination.
(An alternative fix would be to set ignoresdpversion=yes on the peer.)
Sean Bright [Thu, 23 Jan 2020 15:06:35 +0000 (10:06 -0500)]
chan_pjsip: Ignore RTP that we haven't negotiated
If chan_pjsip receives an RTP packet whose payload differs from the
channel's native format, and asymmetric_rtp_codec is disabled (the
default), Asterisk will switch the channel's native format to match
that of the incoming packet without regard to the negotiated payloads.
We now check that the received frame is in a format we have negotiated
before switching payloads which results in these packets being dropped
instead of causing the session to terminate.
George Joseph [Wed, 22 Jan 2020 18:56:38 +0000 (11:56 -0700)]
cdr.c: Set event time on party b when leaving a parking bridge
When Alice calls Bob and Bob does a blind transfer to Charlie,
Bob's bridge leave event generates a finalize on both the party_a
and party_b CDRs but while the party_a CDR has the correct end time
set from the event time, party_b's leg did not. This caused that
CDR's end time to be equal to the answered time and resulted in a
billsec of 0.
* We now pass the bridge leave message event time to
cdr_object_party_b_left_bridge_cb() and set it on that CDR before
calling cdr_object_finalize() on it.
NOTE: This issue affected transfers using chan_sip most of the
time but also occasionally affected chan_pjsip probably due to
message timing.
Sean Bright [Wed, 22 Jan 2020 15:39:47 +0000 (10:39 -0500)]
http: Add ability to disable /httpstatus URI
Add a new configuration option 'enable_status' which allows the
/httpstatus URI handler to be administratively disabled.
We also no longer unconditionally register the /static and /httpstatus
URI handlers, but instead do it based upon configuration.
Behavior change: If enable_static was turned off, the URI handler was
still installed but returned a 403 when it was accessed. Because we
now register/unregister the URI handlers as appropriate, if the
/static URI is disabled we will return a 404 instead.
Additionally:
* Change 'enablestatic' to 'enable_static' but keep the former for
backwards compatibility.
* Improve some internal variable names
Sean Bright [Thu, 16 Jan 2020 19:47:01 +0000 (14:47 -0500)]
app_voicemail, say: Fix various leading whitespace problems
In af90afd90c64c5183c2207d061f9aa15138081b2, Japanese language support
was added to app_voicemail and main/say.c, but the leading whitespace
is not consistent with Asterisk coding guidelines. This patch fixes
that.
Sean Bright [Thu, 16 Jan 2020 15:09:47 +0000 (10:09 -0500)]
chan_sip.c: Stop handling continuation lines after reading headers
lws2sws() does not stop trying to handle header continuation lines
even after all headers have been found. This is problematic if the
first character of a SIP message body is a space or tab character, so
we update to recognize the end of the message header.
Sean Bright [Tue, 14 Jan 2020 22:20:21 +0000 (17:20 -0500)]
app_voicemail: Set globals to default values when voicemail.conf missing
If voicemail.conf exists but is empty, the config parsing process will
default a number of global variables to non-zero values. On the other
hand, if voicemail.conf is missing (arguably semantically equivalent
to an empty file), this process is skipped and the globals are
defaulted to 0.
Set the globals to the same values they would be set to if a
configuration were present. This allows voicemail configuration to be
done completely by Realtime without the need to create an empty
voicemail.conf file.
ASTERISK-27622 #close
Reported by: Jim Van Meggelen
The change in 9b99ef50b5d01ee8111d26efa7b926bdfaf3f980 updated the
syntax of the 'realtime update2' CLI command but did not correctly
update the calls to ast_update2_realtime().
The issue this addresses was originally opened because we aren't
allowing a SQL NULL to be set as part of the update, but this is a
limitation of the Realtime API and is not a bug.
Additionally, this patch:
* Corrects the example in the command documentation to reflect
'update2' instead of 'update.'
* Fixes the leading spacing of the command documentation.
* Checks that the required 'NULL' literal argument is present where we
expect it to be.
Jaco Kroon [Tue, 3 Dec 2019 18:27:38 +0000 (20:27 +0200)]
netsock2: ast_addressfamily_to_sockaddrsize and ast_sockaddr_from_sockaddr.
ast_addressfamily_to_sockaddrize will determine the size that's
required, and ast_sockaddr_from_sockaddr then wraps this new function
and ast_sockaddr_copy_sockaddr to copy arbitrary sockaddr's (without
knowing the address family) into the ast_sockaddr structure.
Kevin Harwell [Wed, 8 Jan 2020 19:54:44 +0000 (13:54 -0600)]
app_agent_pool: Update XML docs for AgentLogin
This patch fixes some wrongly formatted documentation for the AgentLogin
application. A couple of "see also" links should contain only the function
name, and no parameters.
George Joseph [Wed, 8 Jan 2020 18:11:26 +0000 (11:11 -0700)]
CI: Update buildAsterisk.sh to do a "make full"
If you do a "make all" when building Asterisk the xml documentation
produced will be missing certain AMI events where their
documentation is located not at the top of the c source file but
embedded further down next to the event's manager_event()
registration call. See main/manager_mwi.c for an example.
"make full" does produce the correct documentation so we're changing
it in the build script. A separate commit/issue will address the
problem with "make all".
Joshua C. Colp [Mon, 6 Jan 2020 15:02:54 +0000 (15:02 +0000)]
res_pjsip_pubsub: Add ability to persist generator state information.
Some body generators, such as dialog-info+xml, require storing state
information which is then conveyed in the NOTIFY request itself. Up
until now there was no way for such body generators to persist this
information.
Two new API calls have been added to allow body generators to set and
get persisted data. This data is persisted out alongside the normal
persistence information and allows the body generator to restore
state information or to simply use this for normal storage of state.
State is stored in the form of JSON and it is up to the body
generator to interpret this as needed.
The dialog-info+xml body generator has been updated to take advantage
of this to persist the version number.
Sean Bright [Tue, 24 Dec 2019 15:16:23 +0000 (10:16 -0500)]
res_pjsip_endpoint_identifier_ip.c: Add port matching support
Adds source port matching support when IP matching is used:
[example]
type = identify
match = 1.2.3.4:5060/32, 1.2.3.4:6000/32, asterisk.org:4444
If the IP matches but the source port does not, we reject and search for
alternatives. SRV lookups are still performed if enabled (srv_lookups = yes),
unless the configured FQDN includes a port number in which case just a host
lookup is performed.
George Joseph [Mon, 30 Dec 2019 17:04:23 +0000 (10:04 -0700)]
sig_pri: Fix deadlock caused by sig_pri_queue_hangup
The change to add setting hangupsource to sig_pri_queue_hangup()
made in https://gerrit.asterisk.org/c/asterisk/+/12857 casued
deadlocks when a hangup request was received from the core at the
same time a hanguprequest was received from the remote end via the
D channel.
Although the PRI's channel private structure was being unlocked
before setting the hangupsource, the PRI's own lock was still being
held during the process. If channel actions were also coming from
the core, a deadlock on the PRI could result. This deadlock could
then escalate to the entire DAHDI subsystem via DAHDI's global
interface list lock, especially if someone used the PRI CLI commands.
Fix:
* We now unlock the PRI as well as the PRI's channel private
structure before setting the hangupsource, then relock both
afterwards.
Richard Mudgett [Mon, 30 Dec 2019 19:13:46 +0000 (13:13 -0600)]
app_chanisavail.c: Simplify dialplan using ChanIsAvail.
Dialplan has to be careful about passing an empty device list or empty
positions in the list. As a result, dialplan has to check for these
conditions before using ChanIsAvail. Simplify dialplan by making
ChanIsAvail handle these conditions gracefully.
* Made tolerate empty positions in the device list.
* Simplified the code and eliminated some unnecessary indention.
George Joseph [Thu, 2 Jan 2020 20:25:33 +0000 (13:25 -0700)]
stasis.c: Use correct topic name in stasis_topic_pool_delete_topic
When a topic is created for an object, its name is only
<object>:<uniqueid>
For example:
bridge:cb68b3a8-fce7-4738-8a17-d7847562f020
When a topic is added to a pool, its name has the pool's topic
name prepended. For example:
bridge:all/bridge:cb68b3a8-fce7-4738-8a17-d7847562f020
The topic_pool_entry's name however, is only what was passed
in to stasis_topic_pool_get_topic which is
bridge:cb68b3a8-fce7-4738-8a17-d7847562f020
That's actually correct because the entry is qualified by the
pool that's in.
When you're ready to delete the entry from the pool, you retrieve
the tropic name from the object but since it now has the pool's
topic name prepended, it won't be found in the pool container.
Fix:
* Modified stasis_topic_pool_delete_topic() to skip past the
pool topic's name, if it was prepended to the topic name,
before searching the container for a pool entry.
Richard Mudgett [Mon, 30 Dec 2019 04:38:05 +0000 (22:38 -0600)]
app_dial.c: Simplify dialplan using Dial.
Dialplan has to be careful about passing an empty destination list or
empty positions in the list. As a result, dialplan has to check for
these conditions before using Dial. Simplify dialplan by making Dial
handle these conditions gracefully.
* Made tolerate empty positions in the dialed device list.
* Reduced some message log levels from notice to verbose.
Richard Mudgett [Mon, 30 Dec 2019 02:41:30 +0000 (20:41 -0600)]
app_page.c: Simplify dialplan using Page.
Dialplan has to be careful about passing an empty destination list or
empty positions in the list. As a result, dialplan has to check for
these conditions before using Page. Simplify dialplan by making Page
handle these conditions gracefully.
* Made tolerate empty positions in the paged device list.
* Reduced some warnings associated with the 's' option to verbose
messages. The warning level for those messages really serves no purpose
as that is why the 's' option exists.
Richard Mudgett [Mon, 30 Dec 2019 00:36:54 +0000 (18:36 -0600)]
features.c: Make Bridge application tolerate unspecified channel.
The Bridge application was inconsistent if the channel to bridge with is
not specified. If no parameters are given then a warning is issued and
the current channel is hung up. If options are given but no channel is
specified then a warning is issued and the current channel is not hung up.
* Made the Bridge application give a verbose message instead of a warning
if the channel to bridge with is not specified and made not hang up the
current channel. As a result dialplan no longer needs to check if a
channel name is passed before calling Bridge and simply needs to check the
BRIDGERESULT channel variable instead. This is something you likely want
your dialplan to do anyway.
* Fixed up L() option warning message. It is up to the caller to
determine if the channel is hung up because of the warning. Dial() hangs
up the current channel while Bridge() does not.
projects / thirdparty / asterisk.git / log
