Jeff Peeler [Thu, 25 Jun 2009 19:54:12 +0000 (19:54 +0000)]
New signaling module to handle PRI/BRI operations in chan_dahdi
This merge splits the PRI/BRI signaling logic out of chan_dahdi.c into
sig_pri.c. Functionality in theory should not change (mostly). A few trivial
changes were made in sig_analog with verbose messages and commenting.
Russell Bryant [Thu, 25 Jun 2009 16:04:10 +0000 (16:04 +0000)]
Merged revisions 203115 via svnmerge from
https://origsvn.digium.com/svn/asterisk/branches/1.4
........
r203115 | russell | 2009-06-25 11:02:16 -0500 (Thu, 25 Jun 2009) | 11 lines
Resolve a crash related to a T.38 reinvite race condition.
This change resolves a crash observed locally during some T.38 testing.
A call was set up using a call file, and when the T.38 reinvite came in,
the channel state was still AST_STATE_DOWN. The reason is explained by
a comment in the code that previously lived in the handling of
AST_STATE_RINGING. This change modifies the logic to handle the same
race condition for any channel state that is not UP.
* Fixed segfault if the ',' is missing.
* Completely check the range returned by pri_timer2idx() to prevent
possible access outside array bounds.
........
Sean Bright [Wed, 24 Jun 2009 13:47:55 +0000 (13:47 +0000)]
Update sample cdr_tds configuration to try and eliminate some confusion.
Also change the preferred configuration option from 'hostname' (which was
misleading because it didn't actually treat the value as a hostname) to
'connection' and added some verbage explaining that the user would need to
refer to their freetds.conf file for those settings. 'hostname' was kept
as a backwards compatible configuration parameter.
Richard Mudgett [Tue, 23 Jun 2009 21:38:21 +0000 (21:38 +0000)]
Make outgoing_colp=2 misdn.conf port parameter not send redirecting or transfer messages.
If the outgoing_colp parameter is set to not send COLP information, then
it does not make sense to send redirecting or transfer messages announcing
new COLP information that is blocked. The service provider may supply the
listed number for that line when it passes the messages to the next hop.
Why tell the switch that these events happened when the information is
otherwise suppressed?
Also blocked the number of previous redirects that may have occurred to
calls going out the port when outgoing_colp is 2.
David Vossel [Mon, 22 Jun 2009 15:33:35 +0000 (15:33 +0000)]
attempting to load running modules
Modules placed in the priority heap for loading were not properly removed from the linked list. This resulted in some modules attempting to load twice.
Fix a situation in which Asterisk would not stop retransmitting 487s.
If a CANCEL were received by Asterisk, we would send a 487 in response
to the original INVITE and a 200 OK for the CANCEL. If there were a network
hiccup which caused the 200 OK and the 487 to be lost, then the UA communicating
with Asterisk may try to retransmit its CANCEL. Asterisk's response to this used
to be to try sending another 487 to the canceled INVITE and another 200 OK to the
CANCEL.
The problem here is that the originally-sent 487 was sent "reliably" meaning that
it will be retransmitted until it is received properly. So when we receive the second
CANCEL it is likely that the first batch of 487s we sent is still going strong and
reaches the UA. The result was that the second set of 487s would be retransmitted
constantly until the maximum number of retries had been reached.
The fix for this is that if we receive a second CANCEL for an INVITE, then we cancel
the retransmission of the first set of 487s and start a second set. This causes the
dialog to be terminated reasonably.
Fix a possible infinite loop in SDP parsing during glare situation.
There was a while loop in get_ip_and_port_from_sdp which was controlled
by a call to get_sdp_iterate. The loop would exit either if what we were
searching for was found or if the return was NULL. The problem is that
get_sdp_iterate never returns NULL. This means that if what we were searching
for was not present, the loop would run infinitely. This modification of the
loop fixes the problem.
Tilghman Lesher [Fri, 19 Jun 2009 15:47:55 +0000 (15:47 +0000)]
Fix 2 typos and add support for wide character types.
Reported by Benny Amorsen via the asterisk-users mailing list.
http://lists.digium.com/pipermail/asterisk-users/2009-June/233622.html
If the "h" extension fails, give it another chance in main/pbx.c.
If the "h" extension fails, give it another chance in main/pbx.c, when it
returns from the bridge code. Fixes an issue where the "h" extension may
occasionally not fire, when a Dial is executed from a Macro.
Debugged in #asterisk with user tompaw.
........
Tilghman Lesher [Thu, 18 Jun 2009 20:52:36 +0000 (20:52 +0000)]
One of the changes in 1.6.1 was to allow app_directory to use functionality
within app_voicemail for directory functions. It is therefore no longer
necessary for app_directory to be linked against the ODBC libraries (and it
never was necessary for app_directory to be linked against IMAP, though it
was).
Tilghman Lesher [Thu, 18 Jun 2009 18:24:23 +0000 (18:24 +0000)]
Clarify CUT code, and in the process, fix a bug in trunk only
(closes issue #15320)
Reported by: chappell
Patches:
cut_fix.patch uploaded by chappell (license 8)
cut_clarify.patch uploaded by chappell (license 8)
Russell Bryant [Thu, 18 Jun 2009 15:27:10 +0000 (15:27 +0000)]
Merged revisions 201600 via svnmerge from
https://origsvn.digium.com/svn/asterisk/branches/1.4
........
r201600 | russell | 2009-06-18 10:24:31 -0500 (Thu, 18 Jun 2009) | 29 lines
Fix memory corruption and leakage related reloads of non files mode MoH classes.
For Music on Hold classes that are not files mode, meaning that we are executing
an application that will feed us audio data, we use a thread to monitor the
external application and read audio from it. This thread also makes use of the
MoH class object. In the MoH class destructor, we used pthread_cancel() to ask
the thread to exit. Unfortunately, the code did not wait to ensure that the
thread actually went away. What needed to be done is a pthread_join() to ensure
that the thread fully cleans up before we proceed. By adding this one line, we
resolve two significant problems:
1) Since the thread was never joined, it never fully goes away. So, on every
reload of non-files mode MoH, an unused thread was sticking around.
2) There was a race condition here where the application monitoring thread
could still try to access the MoH class, even though the thread executing
the MoH reload has already destroyed it.
Mark Michelson [Thu, 18 Jun 2009 15:20:17 +0000 (15:20 +0000)]
Trunk implementation of setting an alternate RTP source.
This contains the interface by which we can let an rtp instance know
that it might start receiving audio from a new source. This is similar
in nature to revision 197588 of Asterisk 1.4.
David Vossel [Thu, 18 Jun 2009 15:16:05 +0000 (15:16 +0000)]
parsing extension correctly from sip register lines
If a transport type was specified, but no extension, parsing of the extension would return whatever was after the transport rather than defaulting to 's'.
Mark Michelson [Wed, 17 Jun 2009 20:10:01 +0000 (20:10 +0000)]
Fix problem with no audio due to ignoring the SDP.
A recent change to our SDP version comparison made audio not function
on some calls. This was because of a test wherein we were trying to
see if an unsigned value was less than 0. This is a dumb comparison
and arguably the compiler should have warned about it. Alas, though,
it slipped past. Now it's fixed by changing the variable to be a
signed type.
Found by several developers. Tested by mnicholson and dbrooks.
Change the datastore traversal in ast_do_masquerade to use a safe list traversal.
It is possible for datastore fixup functions to remove the datastore from the list
and free it. In particular, the queue_transfer_fixup in app_queue does this. While
I don't yet know of this causing any crashes, it certainly could.
Found while discussing a separate issue with Brian Degenhardt.
........
StopMixMonitor race condition (not giving up file immediately)
StopMixMonitor only indicates to the MixMonitor thread to stop
writing to the file. It does not guarantee that the recording's
file handle is available to the dialplan immediately after execution.
This results in a race condition. To resolve this, the filestream
pointer is placed in a datastore on the channel. When StopMixMonitor
is called, the datastore is retrieved from the channel and the
filestream is closed immediately before returning to the dialplan.
Documentation indicating the use of StopMixMonitor to free files
has been updated as well.
David Vossel [Wed, 17 Jun 2009 15:20:26 +0000 (15:20 +0000)]
SIP registry ref count error
During a sip reload, the list of sip_registry objects are
supposed to be traversed, unlinked, and destroyed, but
destruction never takes place due to a ref counting error.
This causes a memory leak when registry items are removed
from sip.conf and reloaded. While the registries are removed
from the global list, they are not removed from the scheduler.
Because of this, SIP register attempts continue to be sent
out for the item even though it may no longer be in the .conf.
Correct AST_LIST_APPEND_LIST behavior when list to be appended is empty.
When the list to be appended is empty, and the list to be appended to is *not*,
AST_LIST_APPEND_LIST would actually cause the target list to become broken,
and no longer have a pointer to its last entry. This patch fixes the problem.
(reported by Stanislaw Pitucha on the asterisk-dev mailing list)
........
Kevin P. Fleming [Tue, 16 Jun 2009 21:10:15 +0000 (21:10 +0000)]
Enable applications to enable/disable digit and tone detection.
Some applications (notably app_fax) do not need digit detection nor FAX tone
detection while they are running, and if Asterisk is using software DSPs to provide
the detection, this consumes extra CPU cycles that could be better spent on the
actual application. This patch allows applications to query and control the state
of digit and tone detection on a channel, and modifies app_fax to disable them
while the FAX operations are occurring (and re-enable digit detection afterwards).
Kevin P. Fleming [Tue, 16 Jun 2009 20:50:41 +0000 (20:50 +0000)]
When compiling in an Emacs-spawned shell, always print directory names.
This change ensures that Emacs can find the proper source files when parsing
compiler error messages, since it uses the 'make' output including directory
names to do it.
Improve support for media paths that can generate multiple frames at once.
There are various media paths in Asterisk (codec translators and UDPTL, primarily)
that can generate more than one frame to be generated when the application calling
them expects only a single frame. This patch addresses a number of those cases,
at least the primary ones to solve the known problems. In addition it removes the
broken TRACE_FRAMES support, fixes a number of bugs in various frame-related API
functions, and cleans up various code paths affected by these changes.
Kevin P. Fleming [Tue, 16 Jun 2009 16:32:36 +0000 (16:32 +0000)]
Fix problems with new compiler attribute checking in configure script.
The last changes to ast_gcc_attribute.m4 caused some problems checking for
various attributes, because the scope of the symbol the attribute is applied
to can be important; this patch allows the scope to be specified for the check.
David Vossel [Tue, 16 Jun 2009 16:03:30 +0000 (16:03 +0000)]
SIP transport type issues
What this patch addresses:
1. ast_sip_ouraddrfor() by default binds to the UDP address/port
reguardless if the sip->pvt is of type UDP or not. Now when no
remapping is required, ast_sip_ouraddrfor() checks the sip_pvt's
transport type, attempting to set the address and port to the
correct TCP/TLS bindings if necessary.
2. It is not necessary to send the port number in the Contact
header unless the port is non-standard for the transport type.
This patch fixes this and removes the todo note.
3. In sip_alloc(), the default dialog built always uses transport
type UDP. Now sip_alloc() looks at the sip_request (if present)
and determines what transport type to use by default.
4. When changing the transport type of a sip_socket, the file
descriptor must be set to -1 and in some cases the tcptls_session's
ref count must be decremented and set to NULL. I've encountered
several issues associated with this process and have created a function,
set_socket_transport(), to handle the setting of the socket type.
(closes issue #13865)
Reported by: st
Patches:
dont_add_port_if_tls.patch uploaded by Kristijan (license 753)
13865.patch uploaded by mmichelson (license 60)
tls_port_v5.patch uploaded by vrban (license 756)
transport_issues.diff uploaded by dvossel (license 671)
Tested by: mmichelson, Kristijan, vrban, jmacz, dvossel
Michiel van Baak [Tue, 16 Jun 2009 15:51:36 +0000 (15:51 +0000)]
add FILE_STORAGE to Voicemail Build Options
Voicemail can only use one storage module at the moment.
Because it's unclear that selecting one of the storage modules
in menuselect will disable filesystem storage we now have
a FILE_STORAGE option that conflicts with the other modules.
Russell Bryant [Tue, 16 Jun 2009 02:32:33 +0000 (02:32 +0000)]
Don't claim a char * is a mansession object.
Since there was only 1 bucket, and no hash function was specified, the code
actually worked perfectly fine. However, in theory, this was invalid use of
the OBJ_POINTER flag, so remove it so the code provides a better usage example.
Kevin P. Fleming [Tue, 16 Jun 2009 01:28:08 +0000 (01:28 +0000)]
Ensure that configure-script testing for compiler attributes actually works.
The configure script tests for compiler attributes didn't actually enable
enough warnings or provide a proper test harness to determine whether the
compiler supports the attribute in question or not; this caused gcc 4.1 to
report that it supports 'weakref', but it doesn't actually support it in the
way that is needed for our optional API mechanism. The new configure script
test will properly distinguish between full support and partial support
for this attribute, among others.
Kevin P. Fleming [Mon, 15 Jun 2009 20:42:38 +0000 (20:42 +0000)]
Accept T.38 re-INVITE responses with invalid SDP versions.
This commit changes the 'incoming SDP version' check logic a bit more; when
'ignoresdpversion' is *not* set for a peer, if we initiate a re-INVITE to
switch to T.38, we'll always accept the peer's SDP response, even if they
don't properly increment the SDP version number as they should. If this situation
occurs, a warning message will be generated suggesting that the peer's
configuration be changed to include the 'ignoresdpversion' configuration option
(although ideally they'd fix their SIP implementation to be RFC compliant).
Kevin P. Fleming [Mon, 15 Jun 2009 19:10:10 +0000 (19:10 +0000)]
Last batch of 'static' qualifiers for module-level global variables.
Fix up modules in the 'apps' directory, and also correct the bad example of
enum definitions in include/asterisk/app.h, which many developers followed
(thanks for reading the documentation!). In addition, add some basic usage
examples of the 'pahole' and 'pglobal' tools to the coding guidelines.
Kevin P. Fleming [Mon, 15 Jun 2009 16:38:32 +0000 (16:38 +0000)]
Some minor structure size improvements in sip_pvt and sip_peer.
Using the 'pahole' tool, it is now quite easy to see where structure fields
could be organized differently to keep the compiler from having to add
padding to satisfy alignment requirements. These changes reduced the sizes of
sip_pvt and sip_peer by a few bytes each (on 64-bit platforms), and also fixed
a spelling error in a field name.
Kevin P. Fleming [Mon, 15 Jun 2009 16:07:23 +0000 (16:07 +0000)]
Redesigned 'optional API' support.
This patch provides a new implementation of the optional API support defined
in asterisk/optional_api.h; this new version provides solves compatibility
issues with the use of linker version scripts for suppressing global symbols.
In addition, there is now a functional (and tested!) implementation for Mac OS/X,
so module writers no longer need to use special tests before calling optional
API functions. All future implementations must provide these same semantics,
so that module writers can rely on them.
Moises Silva [Sun, 14 Jun 2009 06:13:48 +0000 (06:13 +0000)]
added openr2 to menuselect-deps.in, recent commit in menuselect made me realize this was never done but was working anyways
also added support for skip category request feature of openr2 and updated chan_dahdi.conf.sample
Sean Bright [Fri, 12 Jun 2009 19:42:26 +0000 (19:42 +0000)]
First shot at an upstart script for asterisk on Ubuntu.
This works relatively well (assuming you are using /var/run/asterisk) as your
run directory and upstart 0.3.9. Needs to be generalized and eventually added
to the 'make install' target for Ubuntu.
projects / thirdparty / asterisk.git / log
