Matthew Jordan [Tue, 5 Nov 2013 21:10:20 +0000 (21:10 +0000)]
chan_iax2: Fix incorrect usage of ast_get_ip involving uninitialized struct
This started off as a fix for the failing IAX2 acl_call test in the Asterisk
Test Suite. When inspecting why that test was failing, it became clear that all
attempts to bind to any local loopback address was failing:
[Nov 2 15:56:28] VERBOSE[15787] chan_iax2.c: == Binding IAX2 to address
127.0.0.1:4569
[Nov 2 15:56:28] DEBUG[15787] netsock2.c: Splitting '127.0.0.1' into...
[Nov 2 15:56:28] DEBUG[15787] netsock2.c: ...host '127.0.0.1' and port ''.
[Nov 2 15:56:28] ERROR[15787] netsock2.c: getaddrinfo("127.0.0.1", "(null)",
...): ai_family not supported
[Nov 2 15:56:28] WARNING[15787] acl.c: Unable to lookup '127.0.0.1'
While there's conceivably other ways for getaddrino to return EAI_FAMILY, the
most common way is if AF_INET, AF_INET6, or AF_UNSPEC is not provided as the
desired family. The culprit was the call to ast_get_ip, defined in acl.h. This
function uses the family from the passed in addr object (which it will also
populate when it returns!) when it eventually calls getaddrinfo.
This patch fixes the use of ast_get_ip that were not specifying the family in
chan_iax2. This prevents uninitialized use of the structure, so that the
addresses resolve correctly.
Matthew Jordan [Tue, 5 Nov 2013 21:06:03 +0000 (21:06 +0000)]
netsock2: Define AST_AF_* enum constants to their AF_* equivalents
This patch explicitly defines AST_AF_* enum constants to their sys/socket.h
defined equivalents. It is certainly unclear why these constants actually have
to exist, given that netsock2.h includes sys/socket.h; however, since the code
base is already liberally sprinkled with the usage of AST_AF_* (as well as with
direct calls to AF_*), this will at least keep the semantics consistent between
their usage across systems.
Matthew Jordan [Tue, 5 Nov 2013 20:53:27 +0000 (20:53 +0000)]
stasis_channels: Don't give preference to ANI info in channel snapshots
When publishing channel snapshots, we currently compute the caller ID name and
number by giving preference first to ani.{name|number}, then to
id.{name|number}. However, when a channel driver (such as chan_sip) updates the
caller ID, it typically only updates the caller ID stored in id.{name|number}.
This means that we are currently giving preference to stale information.
When looking at the rest of the code base, the only other place where we appear
to use this same logic is in app_amd. Everywhere else, we treat the party
information in ani as being separate to the party information in id.
This patch publishes only the caller ID name and number in the snapshot field
for caller_name and caller_num. Note that the information in ANI is still
available in caller_ani.
Kevin Harwell [Mon, 4 Nov 2013 20:56:16 +0000 (20:56 +0000)]
chan_sip: notify dialog info ignores presentation indicator in callerid
The presentation indicator in a callerid (e.g. set by dialplan function
Set(CALLERID(name-pres)= ...)) is not checked when SIP Dialog Info Notifies
are generated during extension monitoring. Added a check to make sure the
name and/or number presentations on the callee (remote identity) are set to
allow. If they are restricted then "anonymous" is used instead.
(closes issue AST-1175)
Reported by: Thomas Arimont
Review: https://reviewboard.asterisk.org/r/2976/
........
Merged revisions 402450 from http://svn.asterisk.org/svn/asterisk/branches/11
Richard Mudgett [Sat, 2 Nov 2013 04:05:24 +0000 (04:05 +0000)]
vector: Update API to be more flexible.
Made the vector macro API be more like linked lists.
1) Added a name parameter to ast_vector() to name the vector struct.
2) Made the API take a pointer to the vector struct instead of the struct
itself.
3) Added an element cleanup macro/function parameter when removing an
element from the vector for ast_vector_remove_cmp_unordered() and
ast_vector_remove_elem_unordered().
4) Added ast_vector_get_addr() in case the vector element is not a simple
pointer.
* Converted an inline vector usage in stasis_message_router to use the
vector API. It needed the API improvements so it could be converted.
* Fixed topic reference leak in router_dtor() when the
stasis_message_router is destroyed.
* Fixed deadlock potential in stasis_forward_all() and
stasis_forward_cancel(). Locking two topics at the same time requires
deadlock avoidance.
* Made internal_stasis_subscribe() tolerant of a NULL topic.
* Made stasis_message_router_add(),
stasis_message_router_add_cache_update(), stasis_message_router_remove(),
and stasis_message_router_remove_cache_update() tolerant of a NULL
message_type.
* Promoted a LOG_DEBUG message to LOG_ERROR as intended in
dispatch_message().
Richard Mudgett [Sat, 2 Nov 2013 03:21:17 +0000 (03:21 +0000)]
confbridge: Separate user muting from system muting overrides.
The system overrides the user muting requests when MOH is playing or a
waitmarked user is waiting for a marked user to join. System muting
overrides interfere with what the user may wish the muting to be when the
system override ends.
* User muting requests are now independent of the system muting overrides.
The effective muting is now the logical or of the user request and system
override.
* Added a Muted flag to the CLI "confbridge list <conference>" command.
* Added a Muted header to the AMI ConfbridgeList action ConfbridgeList
event.
Richard Mudgett [Sat, 2 Nov 2013 01:11:16 +0000 (01:11 +0000)]
config: Allow ConfBridge DTMF menus to have '#' as the first digit.
ConfBridge allows custom DTMF menus to be created in the confbridge.conf
file by assigning a DTMF key sequence to a sequence of actions as follows:
DTMF-sequence = action,action...
Unfortunately, the normal config file processing code interprets an
initial '#' character as starting a directive such as #include.
* Add the ability to escape the first non-blank character in a config line
so the '#' character can be used without triggering the directive
processing code.
(closes issue AFS-2)
(closes issue ASTERISK-22478)
Reported by: Nicolas Tanski
Patches:
jira_asterisk_22478_v11.patch (license #5621) patch uploaded by rmudgett (modified)
BridgeDestroy kicks an entire bridge, where BridgeKick kicks just one
channel off the bridge. When kicking a channel, specifying the bridge
also (optional) insures it is not removed from the wrong bridge. The
BridgeTechnology events allow viewing and changing suspension status,
which affects only subsequent not active bridging.
(closes ASTERISK-22356)
Reported by: Richard Mudgett
Review: https://reviewboard.asterisk.org/r/2973/
Joshua Colp [Fri, 1 Nov 2013 14:37:00 +0000 (14:37 +0000)]
res_ari_channels: Add ring operation, dtmf operation, hangup reasons, and tweak early media.
The ring operation sends ringing to the specified channel it is invoked on.
The dtmf operation can be used to send DTMF digits to the specified channel
of a specific length with a wait time in between. Finally hangup reasons
allow you to specify why a channel is being hung up (busy, congestion).
Early media behavior has also been tweaked slightly. When playing media to a channel
it will no longer automatically answer. If it has not been answered a progress indication
is sent instead.
(closes issue ASTERISK-22701)
Reported by: Matt Jordan
Kinsey Moore [Fri, 1 Nov 2013 12:38:22 +0000 (12:38 +0000)]
chan_sip: Fix RTCP port for SRFLX ICE candidates
This corrects one-way audio between Asterisk and Chrome/jssip as a
result of Asterisk inserting the incorrect RTCP port into RTCP SRFLX
ICE candidates. This also exposes an ICE component enumeration to
extract further details from candidates.
(closes issue ASTERISK-21383)
Reported by: Shaun Clark
Review: https://reviewboard.asterisk.org/r/2967/
........
Merged revisions 402345 from http://svn.asterisk.org/svn/asterisk/branches/11
Joshua Colp [Fri, 1 Nov 2013 12:31:51 +0000 (12:31 +0000)]
res_ari_channels: Fix a deadlock when originating multiple channels close to eachother.
If a Stasis application is specified an implicit subscription is done on the originated
channel. This was previously done with the channel lock held which is dangerous as the
underlying code locks the container and iterates items. This change releases the lock
on the originated channel before subscribing occurs.
(closes issue ASTERISK-22768)
Reported by: Matt Jordan
Joshua Colp [Fri, 1 Nov 2013 12:12:32 +0000 (12:12 +0000)]
res_stasis: Ensure the channel is always departed from the bridge when it leaves.
This change adds a command to the command queue to explicitly depart the channel
from the bridge when it is told it has left. If the channel has already been departed
or has entered a different bridge this command will become a no-op.
(closes issue ASTERISK-22703)
Reported by: John Bigelow
(closes issue ASTERISK-22634)
Reported by: Kevin Harwell
Matthew Jordan [Thu, 31 Oct 2013 16:04:59 +0000 (16:04 +0000)]
core/loader: Don't call dlclose in a while loop
For awhile now, we've noticed continuous integration builds hanging on CentOS 6
64-bit build agents. After resolving a number of problems with symbols, strange
locks, and other shenanigans, the problem has persisted. In all cases, gdb
shows the Asterisk process stuck in loader.c on one of the infinite while loops
that calls dlclose repeatedly until success.
The documentation of dlclose states that it returns 0 on success; any other
value on error. It does not state that repeatedly calling it will eventually
clear those errors. Most likely, the repeated calls to dlclose was to force a
close by exhausting the references on the library; however, that will never
succeed if:
(a) There is some fundamental error at work in the loaded library that
precludes unloading it
(b) Some other loaded module is referencing a symbol in the currently loaded
module
This results in Asterisk sitting forever.
Since we have matching pairs of dlopen/dlclose, this patch opts to only call
dlclose once, and log out as an ERROR if dlclose fails to return success. If
nothing else, this might help to determine why on the CentOS 6 64-bit build agent
things are not closing successfully.
Matthew Jordan [Thu, 31 Oct 2013 15:51:36 +0000 (15:51 +0000)]
medix_index: Display errors when library calls fail
Based on feedback from ipengineer in #asterisk, when the media indexer
cannot access a sound file on the system (or otherwise fails) Asterisk
displays a "Cannot frob file" error but fails to tell you why. This is
especially problematic as the media_indexer failing will rpevent Asterisk
from starting, as it is in the core.
We now display the errno error messages so folks can figure out what they've
done wrong.
David M. Lee [Thu, 31 Oct 2013 14:43:44 +0000 (14:43 +0000)]
stasis: add functions embarrassingly missing from r400522
I neglected to implement two of the endpoint subscription functions when
I did the work. Normally, you'll only hit that when you unsubscribe from
a specific endpoint.
Rusty Newton [Tue, 29 Oct 2013 23:43:18 +0000 (23:43 +0000)]
Updates for 1.4.25 core sounds and 1.4.14 extra sounds, plus new en_GB language set
The new sound packages relate to issues: ASTERISK-22544, ASTERISK-22411, ASTERISK-21413, ASTERISK-20782
Modified sounds/Makefile for the new sound versions and to account for the new en_GB language set.
Matthew Jordan [Tue, 29 Oct 2013 12:53:17 +0000 (12:53 +0000)]
Remove some spammy debug messages; improve clarity of others
Debug messages aren't free. Even when the debug level is sufficiently low such
that the messages are never evaluated, there is a cost to having to parse
Asterisk logs that contain debug messages that (a) fail to convey sufficient
information or (b) occur so frequently as to be next to meaningless. Based on
having to stare at lots of DEBUG messages, this patch makes the following
changes:
* channel.c: When copying variables from a parent channel to a child channel,
specify the channels involved. Do not log anything for a variable that is not
inherited; the fact that it doesn't have an _ or __ already signifies that it
won't be inherited.
* pbx.c: Specify what function evaluation has occurred that created the result.
* translate.c: Bump up the translator path messages to 10. I've never once had
to use these debug messages, and for each format that is registered (on
startup) and unregistered (on shutdown) the entire f^2 matrix is logged out.
For short tests in the Asterisk Test Suite, this should make finding the
actual test much easier.
* xmldoc.c: The debug message that 'blah' is not found in the tree is expected.
Often, description elements - which are not required - are not provided.
This debug message adds no additional value, as it is not indicative of an
error or helpful in debugging which element did not contain a 'blah' element
as a child. If an element is supposed to contain a child element, then that
XML tree should have failed validation in the first place.
Kinsey Moore [Tue, 29 Oct 2013 12:51:15 +0000 (12:51 +0000)]
ARI: Remove channels/{channelId}/dial
This removes the /ari/channels/{channelId}/dial URI since it is
redundant, overly complex, is likely to become more externally complex
over time, and is too high-level compared with other ARI operations.
See the following for further information:
http://lists.digium.com/pipermail/asterisk-app-dev/2013-October/000002.html
(closes issue ASTERISK-22784)
Reported by: Matt Jordan
Review: https://reviewboard.asterisk.org/r/2968/
Kinsey Moore [Tue, 29 Oct 2013 12:26:49 +0000 (12:26 +0000)]
bridge_native_rtp: Ensure bridge is torn down
When a bridge transitions away from one tech to another, the tech going
away is provided a dummy bridge with no channels in it to tear down.
Currently this means that the teardown code exits prematurely and does
not tear anything down. This change tears down RTP bridging for the
channel provided in the leave bridge tech callback.
This also reverts the majority of r400403 since it is now redundant.
(closes issue ASTERISK-22628)
(closes issue ASTERISK-22676)
Reported by: John Bigelow
Reported by: Kevin Harwell
Tested by: John Bigelow
Review: https://reviewboard.asterisk.org/r/2905/
Patches:
native_rtp_fix.diff uploaded by Kinsey Moore (License 6273)
Michael L. Young [Mon, 28 Oct 2013 14:51:55 +0000 (14:51 +0000)]
chan_sip: Clarify 'Forcerport' Setting Displayed When Running "sip show peers"
While looking at ASTERISK-22236, Walter Doekes pointed out that when running
"sip show peers", the setting being displayed can be confusing. The display of
"N" used to mean NAT (i.e. yes). The NAT setting has gone through many
different changes resulting in the display of different characters to try and
convey what the current setting is for 'Forcerport' (A for Auto and Forcerport
is currently on, a for Auto but Forcerport is off, Y for yes, and N for no).
During the initial code review to try and clarify these settings (especially
since "N" no longer meant what it used to mean in prior versions of Asterisk),
Mark Michelson suggested using the full space available to display the settings
which helped to make the settings very clear. That was a great suggestion.
Therefore, this patch does the following:
* The column for 'Forcerport' now will show: Auto (Yes), Auto (No), Yes, or No.
* A column for the 'Comedia' setting has been added. It too will display the
setting in a non-cryptic way: Auto (Yes), Auto (No), Yes, or No.
* UPGRADE.txt has been updated to document this change.
(closes issue ASTERISK-22728)
Reported by: Walter Doekes
Tested by: Michael L. Young
Patches:
asterisk-forcerport-display-clarification_v3.diff
uploaded by Michael L. Young (license 5026)
Matthew Jordan [Sun, 27 Oct 2013 23:22:19 +0000 (23:22 +0000)]
Filter out internal channels from dial message handling
Surrogate channels would pop up from time to time in dial message handling.
This would cause a WARNING message to appear, indicating that the Surrogate
channel had no CDR. This patch filters out those channels that have the
internal implementation flag set, such that the WARNING message isn't
displayed.
Matthew Jordan [Sun, 27 Oct 2013 19:40:43 +0000 (19:40 +0000)]
Prevent CDR backends from unregistering while billing data is in flight
This patch makes it so that CDR backends cannot be unregistered while active
CDR records exist. This helps to prevent billing data from being lost during
restarts and shutdowns.
rtp_engine: fix rtp payloads copy and improve argument names
In function ast_rtp_instance_early _bridge_make_compatible the
use of instance 0/1 as arguments doesn't clearly communicate a
direction that the copying of payloads from the source channel
to the destination channel will occur, making it more probable
to have the arguments to ast_rtp_codecs_payloads_copy() put in
the reverse order. This patch renames the arguments with _dst
and _src suffixes and corrects the copy direction.
(closes issue ASTERISK-21464)
Reported by: Kevin Stewart
Review: https://reviewboard.asterisk.org/r/2894/
........
Merged revisions 402000 from http://svn.asterisk.org/svn/asterisk/branches/1.8
Test shows rtpmap:119 being copied per this change, but is not in sip invite
........
Merged revisions 402042 from http://svn.asterisk.org/svn/asterisk/branches/11
Jonathan Rose [Fri, 25 Oct 2013 21:20:42 +0000 (21:20 +0000)]
ARI recordings: Issue HTTP failures for recording requests with file conflicts
If a file already exists in the recordings directory with the same name as what
we would record, issue a 422 instead of relying on the internal failure and
issuing success.
pbx.c: fix confused match caller id that deleted exten still in hash
This fixes a bug where a zero length callerid match adjacent to a no
match callerid extension entry would be deleted together, which then
resulted in hashtable references to free'd memory. A third state of
the matchcid value has been added to indicate match to any extension
which allows enforcing comparison of matchcid on/off without errors.
Kevin Harwell [Fri, 25 Oct 2013 16:07:38 +0000 (16:07 +0000)]
chan_sip: Allow a sip peer to accept both AVP and AVPF calls
Adapts the behaviour of avpf to only impact the format of outgoing calls. For
inbound calls, both AVP and AVPF calls will be accepted regardless of the value
of avpf in the configuration.
David M. Lee [Fri, 25 Oct 2013 13:48:40 +0000 (13:48 +0000)]
test_json: Fix deprecation warnings
After a series of upgrades over recent weeks, I've discovered that
test_json.c won't compile in dev mode any more for me.
One of gcc-4.8.2, OS X Mavericks or Xcode 5 has decided to deprecate
tempnam. Which, in general, is a good thing. But for test code that just
needs a temporary file, it's just annoying.
This patch replaces usage of tempname with mkstemp, avoiding the
deprecation warning. It also removes the temporary files when the test
is complete, which apparently we weren't doing before (oops).
Kevin Harwell [Thu, 24 Oct 2013 20:56:50 +0000 (20:56 +0000)]
Logging: Logging types ignored after specifying a verbose level
If one specified a verbose level within a logging facility in
logger.conf then any component after it was ignored. Fixed so
all values are correctly read.
(closes issue ASTERISK-22456)
Reported by: Kevin Harwell
........
Merged revisions 401833 from http://svn.asterisk.org/svn/asterisk/branches/11
David M. Lee [Thu, 24 Oct 2013 03:12:06 +0000 (03:12 +0000)]
The Swagger 1.2 specification for type extension ended up being
slightly different than my proposal. Instead of putting an 'extends'
field on the subtype, the base type has a 'subTypes' field, which is a
list of the subTypes. Given that its a messaging model and not an
object model, kinda makes sense.
This patch changes the events.json api-doc, and the python translators
to take the new format into account.
Other changes that are in Swagger 1.2 were not adopted, since the spec
is still in flux, and could change before it's finalized.
A summary of changes to the Swagger-1.2 spec can be found at
https://github.com/wordnik/swagger-core/wiki/1.2-transition.
Joshua Colp [Wed, 23 Oct 2013 11:14:49 +0000 (11:14 +0000)]
chan_sip: Fix an issue where an incompatible audio format may be added to SDP.
If preferred codecs included any non-audio format the code would
mistakenly add the audio format, even if it was not a joint capability
with the remote side.
Michael L. Young [Wed, 23 Oct 2013 02:31:48 +0000 (02:31 +0000)]
chan_iax2: Fix Binding To Multiple Addresses Again
When reworking chan_iax2 for IPv6, the ability to bind to multiple addresses
was removed by mistake. This patch restores this functionality and adds notes
about IPv6 addresses in the sample config.
(closes issue ASTERISK-22741)
Reported by: Joshua Colp
Tested by: Michael L. Young
Patches:
asterisk-22741-fix-binding-multiple-addr.diff
uploaded by Michael L. Young (license 5026)
Matthew Jordan [Tue, 22 Oct 2013 22:50:23 +0000 (22:50 +0000)]
res_rtp_asterisk: Fix crash when RTCP is not available during SSRC change
In r400089, a patch was put in to correct erroneous RTCP statistic resets.
Unfortunately, ast_rtp_read can be called on an RTP instance that does not
have RTCP information. This patch prevents that crash by only resetting
the statistics if we do actually have an RTCP instance.
(issue AST-1174)
(closes issue ASTERISK-22667)
Reported by: John Bigelow
........
Merged revisions 401445 from http://svn.asterisk.org/svn/asterisk/branches/1.8
........
Merged revisions 401446 from http://svn.asterisk.org/svn/asterisk/branches/11
The queue_log entry resulting from CLI "queue remove member" when
log_membername_as_agent is enabled is wrong. It always uses the interface
name instead of the member name in the queue_log entry.
* Get the queue member before removing it from the queue so the member
name is available for the queue_log entry.
(closes issue ASTERISK-21826)
Reported by: Oscar Esteve
Patches:
fix_membername.diff (license #6505) patch uploaded by Oscar Esteve
(modified to fix potential ref leak)
........
Merged revisions 401433 from http://svn.asterisk.org/svn/asterisk/branches/11
Richard Mudgett [Tue, 22 Oct 2013 17:05:14 +0000 (17:05 +0000)]
Bridging: Fix orphaned bridge if neither of the joining channels can join.
The original issue noted that the bridge is orphaned when res_parking.so
is not loaded and a call uses the dial kK flags.
A similar issue happens when only one of the park flags is used. In this
case you have the bridge with one or the other channel left in it. The
channel and bridge will stay around until the channel hangs up.
* Fixed the initial bridge channel push failure to act as if the channel
were kicked out of the bridge. The bridge then decides if it needs to be
dissolved.
(closes issue ASTERISK-22629)
Reported by: Kevin Harwell
Richard Mudgett [Tue, 22 Oct 2013 16:32:23 +0000 (16:32 +0000)]
res_parking: Give parking timeout comebacktoorigin channel DTMF features.
Parking timeouts did not set any DTMF features for the channel calling the
parker back.
* Added code to set the parkedcalltransfers, parkedcallreparking,
parkedcallhangup, and parkedcallrecording options appropriately for the
channels when a parking timeout occurs. The recall channel DTMF options
are set using the BRIDGE_FEATURES channel variable to allow the other
timeout options to have the DTMF features available.
(closes issue ASTERISK-22630)
Reported by: Kevin Harwell
Richard Mudgett [Tue, 22 Oct 2013 16:26:20 +0000 (16:26 +0000)]
res_parking: Update XML documention for DTMF features after parking timeout.
* Updated the XML documentation to indicate that the parkedcalltransfers,
parkedcallreparking, parkedcallhangup, and parkedcallrecording
configuration options also apply to parking timeouts.
Mark Michelson [Mon, 21 Oct 2013 21:05:37 +0000 (21:05 +0000)]
Remove a noisy debug message from bridging code.
This particular debug message, during a stress test, was logged so
often that it appeared that there may be a memory leak in the logger
code. In actuality, there was no memory leak, but the logger thread
was having a hard time keeping up with the demands of the rest of the
system.
Since this debug message has no value at all, the best way to fix the
problem was to just remove the message.
(closes issue AST-1225)
reported by John Bigelow
Patches:
spammy_log.diff uploaded by Mark Michelson (License #5049)
Kinsey Moore [Sat, 19 Oct 2013 21:53:08 +0000 (21:53 +0000)]
Fix IAX2 incoming call address lookups
This fixes address lookup for incoming calls without a peer definition.
The address family was unset instead of being set to AST_AF_UNSPEC
which was causing lookup failures on "127.0.0.1". This is one of the
causes of the current failure of the app_page integration test.
Joshua Colp [Sat, 19 Oct 2013 14:43:41 +0000 (14:43 +0000)]
Return a channel snapshot when originating using ARI, and subscribe the Stasis application to it.
This change allows a user of ARI to know what channel it has originated and also follow any
progress. If a Stasis application is provided it will be automatically subscribed to the
originated channel immediately.
(closes issue ASTERISK-22485)
Reported by: David Lee
David M. Lee [Fri, 18 Oct 2013 21:49:48 +0000 (21:49 +0000)]
This is just a quick script for dumping swagger-ui into static-http,
so that it can be served by the Asterisk web server.
I had to change the Makefile in order to recursively install content
from the static-http directory, hence the code review instead of just
putting it in.
This is nice because it automatically takes care of the object references
for you. However, there is a pitfall here. If a break statement is in the
for loop, then the current reference is not cleaned up. In some cases, this
is on purpose, but in others there is a leak. This commit fixes the leak
cases.
Richard Mudgett [Fri, 18 Oct 2013 16:52:32 +0000 (16:52 +0000)]
Add channel lock protection around translation path setup.
Most callers of ast_channel_make_compatible() happen before the channels
enter a two party bridge. With the new bridging framework, two party
bridging technologies may also call ast_channel_make_compatible() when
there is more than one thread involved with the two channels.
* Added channel lock protection in set_format() and
ast_channel_make_compatible_helper() when dealing with the channel's
native formats while setting up a translation path.
* Fixed best_src_fmt and best_dst_fmt usage consistency in
ast_channel_make_compatible_helper(). The call to
ast_translator_best_choice() got them backwards.
* Updated some callers of ast_channel_make_compatible() and the function
documentation. There is actually a difference between the two channels
passed in.
* Fixed the deadlock potential in res_fax.c dealing with
ast_channel_make_compatible(). The deadlock potential was already there
anyway because res_fax called ast_channel_make_compatible() with chan
locked.
(closes issue ASTERISK-22542)
Reported by: Matt Jordan
Michael L. Young [Fri, 18 Oct 2013 15:13:02 +0000 (15:13 +0000)]
Remove Port Restriction When Checking For NAT
When trying to determine if a peer is behind NAT, we should not be using the
ports when comparing addresses.
This patch removes the port from being checked and just useds the addresses
now.
(closes issue ASTERISK-22729)
Reported by: Michael L. Young
Tested by: Michael L. Young
Patches:
asterisk-remove-using-port-for-nat-check.diff
uploaded by Michael L. Young (license 5026)
Walter Doekes [Fri, 18 Oct 2013 14:47:12 +0000 (14:47 +0000)]
Properly copy/remove the device state cache flag over a masquerade.
In r378303 the AST_FLAG_DISABLE_DEVSTATE_CACHE flag was added that tells
the devstate system to not cache states for non-real devices. However,
when optimizing away channels (ast_do_masquerade), that flag wasn't
copied.
In my case, using Local devices as queue members created a situation
where the endpoint was considered in use, but the state change of the
device being available again was ignored (not cached). The endpoint
channel was optimized into the (previously) Local channel, but kept
the do-not-cache flag. The end result being that the queue member
apparently stayed in use forever.
(closes issue ASTERISK-22718)
Reported by: Walter Doekes
Michael L. Young [Thu, 17 Oct 2013 20:37:10 +0000 (20:37 +0000)]
Fix Setting A chan_sip Dialog's SIP_NAT_FORCE_RPORT Flag
A condition was added in a commit to fix ASTERISK-21374, that, if the
SIP_PAGE3_NAT_AUTO_RPORT flag was set, to then copy a peer's SIP_NAT_FORCE_RPORT
flag to the dialog. This condition should not have been there since it assumed
that if Asterisk is in an environment where NAT is involved, that the auto_* nat
settings or force_rport setting would be on in the global settings. If the nat
setting in the global setting is set to 'nat=no' and then turned on for peers
(which is not quite the recommended way, although it is allowed) this flag is
never copied to the dialog resulting in problems like, REGISTER replies going
to the wrong port.
This patch removes this conditional check and will now always use the peer's
flag which by this point in the code the checks on whether the peer is behind
NAT or not (if using auto_force_rport) have already been run.
(closes issue ASTERISK-22236)
Reported by: Filip Frank
Tested by: Michael L. Young
Patches:
asterisk-2236-always-set-rport.diff uploaded
by Michael L. Young (license 5026)
Kinsey Moore [Wed, 16 Oct 2013 14:01:04 +0000 (14:01 +0000)]
Clarify documentation for channel and bridge list
This makes it clear that the ARI API calls for listing channels and
bridges will list all channels or bridges in the system and not just
those that are in or are controlled by a Stasis application.
(closes issue ASTERISK-22635)
Reported by: Kevin Harwell
Walter Doekes [Wed, 16 Oct 2013 12:12:42 +0000 (12:12 +0000)]
Don't check all realtime queues when doing "queue show some_queue".
When using realtime queues, queues have to be fetched from the database
every now and then to see if any info has been changed or to see if the
queue has been removed. When fetching info for an individual queue, the
pruning of other queues is unnecessarily costly.
Kinsey Moore [Tue, 15 Oct 2013 20:02:08 +0000 (20:02 +0000)]
Ensure bridge record error responses validate
This adds the list of expected errors to the /bridges/{bridgeId}/record
ARI documentation so that outbound 4xx errors validate properly.
Previously, this would result in a response validation failure.
Mark Michelson [Tue, 15 Oct 2013 15:21:56 +0000 (15:21 +0000)]
Prevent chan_sip from sending duplicate BYEs.
When a 200 OK for an initial INVITE is received, we were doing
the right thing by ACKing and sending an immediate BYE. However,
we also were doing the wrong thing and queuing an answer frame,
thus causing the call to be answered. This would cause the call
to be hung up by the channel thread, thus resulting in a second
BYE being sent out.
In this fix, I also have set the hangupcause to be correct since
the initial BYE being sent by Asterisk had an unknown hangup
cause. I have changed to using "Bearer capabilty not available"
since the call was hung up due to an SDP offer/answer error.
(closes issue ASTERISK-22621)
reported by Kinsey Moore
........
Merged revisions 400970 from http://svn.asterisk.org/svn/asterisk/branches/1.8
........
Merged revisions 400971 from http://svn.asterisk.org/svn/asterisk/branches/11
David M. Lee [Tue, 15 Oct 2013 13:43:05 +0000 (13:43 +0000)]
My doc correction in r400842 had a silly bug.
Because I added a wiki_description to models and not their properties, the
rendered wiki page had the model description instead of the property
descriptions, which looks very silly indeed.
Mark Michelson [Mon, 14 Oct 2013 21:52:24 +0000 (21:52 +0000)]
chan_sip: Do not increment the SDP version between 183 and 200 responses.
Bumping the SDP version number can cause interoperability problems
since receivers of the responses will expect that a 200 SDP will
be identical to a previous 183 SDP.
(closes issue ASTERISK-21204)
reported by NITESH BANSAL
Patches:
dont-increment-session-version-in-2xx-after-183.patch uploaded by NITESH BANSAL (License #6418)
........
Merged revisions 400906 from http://svn.asterisk.org/svn/asterisk/branches/1.8
........
Merged revisions 400908 from http://svn.asterisk.org/svn/asterisk/branches/11
Kevin Harwell [Mon, 14 Oct 2013 15:52:28 +0000 (15:52 +0000)]
pjsip outbound registration: Log message says received a 408 when we didn't
If the server didn't exist that we are trying to register to the log message
would say that a 408 was received from that server when in reality one wasn't.
Added log messages stating no response was received if the response does not
exist.
(closes issue ASTERISK-22554)
Reported by: Rusty Newton
Review: https://reviewboard.asterisk.org/r/2893/
Joshua Colp [Sun, 13 Oct 2013 15:41:37 +0000 (15:41 +0000)]
Fix a race condition in res_pjsip_session with rapidly terminating the session.
The INVITE session state callback wrongly assumes that a session will always exist, but
when rapidly terminating the session this assumption goes out the window. As all handler
code for the INVITE session state callback requires the session it will now just exit
immediately if no session exists.
(closes issue ASTERISK-22668)
Reported by: John Bigelow
Kinsey Moore [Sat, 12 Oct 2013 16:49:00 +0000 (16:49 +0000)]
Fix realm comparison for outbound auth
When generating the list of authentication credentials to pass to
PJSIP, Asterisk was using the raw pointer of a pj_str_t which is not
always NULL-terminated. This sometimes resulted in incorrect text for
the realm and a failure to match the realm for authentication purposes
which was causing the outbound nominal auth pjsip basic call test to
bounce. This now uses the pj_str_t that contains the realm instead of
generating a new one. Thanks to John Bigelow for helping to narrow this
down.
projects / thirdparty / asterisk.git / log
