Kinsey Moore [Tue, 19 Jul 2011 15:43:32 +0000 (15:43 +0000)]
MeetMe requests a PIN twice in some circumstances
If a call to MeetMe includes both the dynamic(D) and always request PIN(P)
options, MeetMe will ask for the PIN two times: once for creating the
conference and once for entering the conference. This behavior was introduced
in rev 311616 when adding the CONFFLAG_ALWAYSPROMPT option to the logic branch
controlling PIN entry for joining a conference.
Mark Murawki [Mon, 18 Jul 2011 20:47:04 +0000 (20:47 +0000)]
app_dial may double free a channel datastore
When starting a call with originate, and having the callee channel run Bridge() on pickup, we will double free the dialed_interface_info datastore, causing a crash. Make sure to check if the datastore still exists before trying to free it.
(closes issue ASTERISK-17917)
Reported by: Mark Murawski
Tested by: Mark Murawski
Alexandr Anikin [Fri, 15 Jul 2011 19:22:24 +0000 (19:22 +0000)]
small gk processing fixes:
- decrease for 1 second registration ttl for very low expirations (some
providers expire few earlier than TTL)
- delete rrq and registration expire timers on URQ received as we make
new registration.
Introduce <support_level> tags in MODULEINFO.
This change introduces MODULEINFO into many modules in Asterisk in order to show
the community support level for those modules. This is used by changes committed
to menuselect by Russell Bryant recently (r917 in menuselect). More information about
the support level types and what they mean is available on the wiki at
https://wiki.asterisk.org/wiki/display/AST/Asterisk+Module+Support+States
Kevin P. Fleming [Tue, 12 Jul 2011 22:53:53 +0000 (22:53 +0000)]
Correct double-free situation in manager output processing.
The process_output() function calls ast_str_append() and xml_translate() on its
'out' parameter, which is a pointer to an ast_str buffer. If either of these
functions need to reallocate the ast_str so it will have more space, they will
free the existing buffer and allocate a new one, returning the address of the
new one. However, because process_output only receives a pointer to the ast_str,
not a pointer to its caller's variable holding the pointer, if the original
ast_str is freed, the caller will not know, and will continue to use it (and
later attempt to free it).
Matthew Jordan [Tue, 12 Jul 2011 19:10:34 +0000 (19:10 +0000)]
Added additional checks for mailbox / password beginning with '*' character
A bug existed such that if a user entered a password with '*', and the extension 'a' did not exist, an invalid mailbox would be created and the user authenticated. The code was changed to prevent this from occurring, and to prevent users from having mailboxes or passwords defined that begin with the '*' character.
(closes issue ASTERISK-17443)
Reported by: Kevin Scott Adams
Tested by: Matt Jordan
Use 'printf' (POSIX issue 4) instead of 'echo -n', for portability.
The problem with using 'echo -n' is that it is not portable. While BSD systems
required that the '-n' option be removed and interpreted, System V required
that all strings should be echoed with no interpretation of options. This
fundamental difference of behavior means that it is never possible to use the
'-n' flag to echo in tests which are meant to be portable.
In this case, on Mac OS X 10.6, the /bin/sh shell builtin 'echo' uses the
System V semantics of the command, and thus the SHELL test failed on that
platform.
Terry Wilson [Mon, 11 Jul 2011 19:41:59 +0000 (19:41 +0000)]
Update chan_gtalk to work with changed GMail-based calls
The messages sent by the GMail client have changed, but include the
old-style messages as well. This patch checks for this case and
uses the old-style offer.
Richard Mudgett [Fri, 8 Jul 2011 21:41:58 +0000 (21:41 +0000)]
INVITE 403 Forbidden response always retransmits the maximum times.
Asterisk sends a 403 Forbidden response if authentication fails for an
INVITE as required. However, it ignores the ACK and keeps retransmitting
the response.
* Made not delete the to-tag in the dialog so the expected ACK can be
matched with the dialog and stop the retransmissions.
Reset our ast_str before passing it on to dialplan function backends.
It is possible for a dialplan backend to not modify the given buffer or ast_str
and still return success. This causes any previous value stored in the buffer
to be used as if the new function call provided it. Some functions also append
to the given buffer assuming it is empty.
The test_substitution unit test has also been modified to detect this problem.
Jonathan Rose [Thu, 7 Jul 2011 16:04:51 +0000 (16:04 +0000)]
res_odbc patch by tilghman to fix integers with null values
Addresses some improper sql statements in res_odbc that would cause an update to fail on
realtime peers due to trying to set as "(NULL)" rather than an actual NULL.
David Vossel [Wed, 6 Jul 2011 15:26:49 +0000 (15:26 +0000)]
Reverts fix for timerfd locking issue.
jrose discovered a performance issue with this
fix that prevents his analog phones from working
when using timerfd as a timing source. Until
it is understood what is causing this performance
problem, this patch is being reverted.
Add the attribute "type" to each "<use>" for menuselect.
This matters only when autoconf fails to detect that weak linking is supported.
External optional dependencies will become optional in both cases, as they are
removed at compile time when not detected. However, runtime-optional modules
are made mandatory when weak linking is not found. This change affects only
the external optional dependencies; previously, they were incorrectly required
when weak linking support was not detected.
Patches:
20110702__issue18062__asterisk_trunk.diff.txt by tilghman (License #5003)
Matthew Jordan [Tue, 5 Jul 2011 13:23:57 +0000 (13:23 +0000)]
Updated filestream destructor to block until move is complete when cache is used
When a cache directory is used, the process is forked and a mv command is executed to move the temporary file to the permanent location. This caused issues with voicemail, where a race condition occurred when the parent expected the file to be in the permanent location prior to the mv command completing. The parent process is now blocked until the mv command completes.
(closes issue ASTERISK-17724)
Reported by: Adiren P.
Tested by: mjordan
Richard Mudgett [Fri, 1 Jul 2011 21:07:22 +0000 (21:07 +0000)]
Better way to get chan and pvt lock for issue ASTERISK-17431.
Redoes -r308945 for issue ASTERISK-17431 deadlock fix for
sip_set_udptl_peer() and sip_set_rtp_peer().
* Lock the channels in the defined order and avoid the need for a deadlock
avoidance loop.
* Lock the channel before getting the pointer to the private structure to
be sure that the pointer will not change due to a masquerade or channel
hangup.
* To preserve sanity, check that chan and p->owner are the same. (Pointer
rearangements should not happen without the protection of locks because
bad things tend to happen otherwise.)
Kinsey Moore [Wed, 29 Jun 2011 21:49:21 +0000 (21:49 +0000)]
chan_sip: cleanup from the introduction of ast_str
Remove the length field from sip_req and sip_pkt in chan_sip since they are
redundant (ast_str holds its own length) and refactor the necessary functions.
Richard Mudgett [Wed, 29 Jun 2011 18:05:15 +0000 (18:05 +0000)]
Response to QueueRule manager command does not contain ActionID if it was specified.
* Add ActionID support as documented for the QueueRule AMI action.
* Remove documentation for ActionID with the Queues AMI action. The
output does not follow normal AMI response output and there is no place to
put an ActionID header.
Richard Mudgett [Mon, 27 Jun 2011 15:37:19 +0000 (15:37 +0000)]
When subscribing MWI to an unsolicited mailbox the first notification is incorrect.
A remote peer subscribed to MWI with the unsolicited option and a local
phone subscribed to the remote mailbox. The notify message-summary events
are sent correctly except for the first one when subscribing, which will
always be 0. This means the phone MWI indicator will be wrong until the
mailbox read/unread count changes and the event is fired.
Looks like this is a regression from ASTERISK-16149.
* Fix the logic to check the cache and if allowed then fallback to
manually counting mailbox messages.
Richard Mudgett [Fri, 24 Jun 2011 20:46:01 +0000 (20:46 +0000)]
Syntax errors in dialplan do not display the file name.
When issuing the CLI command "dialplan reload" syntax errors and warnings
are displayed on the console. The offending line number is displayed on
the console, but the file name is not displayed. Errors caught in
main/config.c do display the file name.
(closes issue ASTERISK-17985)
Reported by: ulogic
Patches:
pbx_config.patch uploaded by ulogic (License #5685) modified format
Tested by: rmudgett
Jonathan Rose [Fri, 24 Jun 2011 16:48:06 +0000 (16:48 +0000)]
DTMF wasn't being logged on connected consoles when enabled in logger.conf
Previously in order for DTMF to be logged in a connected console session, the user would
have to do logger set channel DTMF on. This corrects that so that it is on by default.
This issue was caused by an off by one error incurred by a logger level count of 6 in
logger.h where it should have been 7.
(closes issue: ASTERISK-17974)
Reported by: Luke H
Terry Wilson [Wed, 22 Jun 2011 18:52:04 +0000 (18:52 +0000)]
Stop sending IPv6 link-local scope-ids in SIP messages
The idea behind the patch listed below was used, but in a more targeted manner.
There are now address stringification functions for addresses that are meant to
be sent to a remote party. Link-local scope-ids only make sense on the machine
from which they originate and so are stripped in the new functions.
There is also a host sanitization function added to chan_sip which is used
for when peer and dialog tohost fields or sip_registry hostnames are used to
craft a SIP message.
Also added are some basic unit tests for netsock2 address parsing.
Richard Mudgett [Wed, 22 Jun 2011 18:41:20 +0000 (18:41 +0000)]
Timout or error on INFO or MESSAGE transaction causes call to be lost.
When exchanging INFO messages within a call, 4xx error causes the call to
be disconnected although RFC 2976 explicitly states that such transactions
do not modify the state of the dialog.
When exchanging MESSAGE messages within a call, 4xx error causes the call
to be disconnected. To provide least surprise, we should not disconnect
the call since a MESSAGE is like INFO in this case. (Implied by RFC 3428
Section 2)
David Vossel [Tue, 21 Jun 2011 20:11:52 +0000 (20:11 +0000)]
Fixes locking inversion issue in ast_async_goto()
During this function we can not hold the "chan" lock while
doing the masquerade, the explicit goto on the tmp chan, or
the channel alloc. Instead we need to get the channel lock,
store off information about the channel that we need, and
then let the channel lock go for the remainder of the function.
Kinsey Moore [Tue, 21 Jun 2011 16:09:14 +0000 (16:09 +0000)]
ConfBridge does not handle hangup properly
When playing back a prompt to a channel, confbridge neglects to check for
hangup events causing lockup condititions for hangups that occur before
actually joining the conference. This change ensures that the user is removed
from the conference in the event of a premature hangup.
Terry Wilson [Thu, 16 Jun 2011 22:35:41 +0000 (22:35 +0000)]
Lock the channel before calling the setoption callback
The channel needs to be locked before calling these callback functions. Also,
sip_setoption needs to lock the pvt and a check p->rtp is non-null before using
it.
DYNAMIC_FEATURES were broken by a recent DTMF change. This patch makes
sure that dynamic features are also checked when deciding whether or not
to pass DTMF through or store it for interpreting.
Jonathan Rose [Wed, 15 Jun 2011 17:42:42 +0000 (17:42 +0000)]
Adds locking to find_table in res_configure_pgsql to prevent a crash.
Bryonclark described the problem as occuring during this function because of multiple
simultaneous database operations causing corruption against a pgsqlConn object.
Richard Mudgett [Wed, 15 Jun 2011 16:43:18 +0000 (16:43 +0000)]
[regression] Voicemail MWI is no longer sent.
When leaving a voicemail, the MWI message is never sent. The same thing
happens when checking a voicemail and marking it as read.
If you restart Asterisk, everything comes up at that state correctly, but
changes to the messages in voicemail causes the light to not be set
appropriately. Very easy to reproduce.
* Made ast_event_check_subscriber() return TRUE if there are ANY
subscribers to an event type when there are no restricting ie values
passed. This allows an event being queued to be queued.
Resolve a segfault/bus error when we try to map memory that falls on a page
boundary.
The fix for ASTERISK-15359 was incorrect in that it added 1 to the length of the
mmap'd region. The problem with this is that reading/writing to that extra byte
outside of the bounds of the underlying fd causes a bus error.
The real issue is that we are working with both a FILE * and the raw fd
underneath it and not synchronizing between them. The code that was removed in
ASTERISK-15359 was correct, but we weren't flushing the FILE * before mapping
the fd.
Looking at the manager code in 1.4 reveals that the FILE * in 'struct
mansession' is never used except to create a temporary file that we immediately
fdopen. This means we just need to write a 0 byte to the fd and everything will
just work. The other branches require a call to fflush() which, while not a
guaranteed fix, should reduce the likelihood of a crash.
This all makes sense in my head.
(closes issue ASTERISK-16460)
Reported by: Ravelomanantsoa Hoby (hoby)
Patches:
issue17747_1.4_svn_markII.patch uploaded by Sean Bright (license #5060)
........
................
Jonathan Rose [Tue, 14 Jun 2011 16:38:43 +0000 (16:38 +0000)]
Changes contact use in build_peer to use the FORCE_RPORT flag instead of RPORT_PRESENT
It turned out that this was causing NAT=Yes to always use rport when present which was
against 1.6.2 behavior and the check itself was redundant since the only way this
segment of code could be reached was if RPORT_PRESENT was already evaluated as true
earlier.
Leif Madsen [Mon, 13 Jun 2011 20:22:21 +0000 (20:22 +0000)]
Additional documentation for bindaddr.
Note that bindaddr will only enable UDP instead of both UDP and TCP which is
what I would expect for backwards compatibility with systems being upgraded
which only support UDP transportation.
(closes issue ASTERISK-17976)
Reported by: Sean Darcy
Terry Wilson [Fri, 10 Jun 2011 15:29:00 +0000 (15:29 +0000)]
Avoid a DB1 infinite loop bug
Explicity check the last entry in the DB and make sure that we don't iterate
past it. Since there can be no duplicates, this just makes sure that we stop
after matching the last key.
This patch also refactors the code to get away from some code duplication. A
previous patch added many astdb tests and this patch passed them.
projects / thirdparty / asterisk.git / log
