* Classes are now tracked past removal from the core container, and module
removal is actively prevented until all references are freed.
* A hanging reference stored in the channel has been removed. This could have
caused a mismatch and the music state not properly cleared, if two or more
reloads occurred between MOH being stopped and MOH being restarted.
* In certain circumstances, duplicate classes were possible.
* A race existed at reload time between a process being killed and the thread
responsible for reading from the related pipe respawning that process.
* Several reference counts have also been corrected. At least one could have
caused deleted classes to stick around forever, consuming resources. This
originally manifested as MOH external processes that were not killed at
reload time.
(closes issue #16279, closes issue #16207)
Reported by: parisioa, dcabot
Patches:
20091202__issue16279__2.diff.txt uploaded by tilghman (license 14)
Tested by: parisioa, tilghman
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@232660 65c4cc65-6c06-0410-ace0-fbb531ad65f3
https://origsvn.digium.com/svn/asterisk/branches/1.4
........
r222152 | kpfleming | 2009-10-05 20:16:36 -0500 (Mon, 05 Oct 2009) | 20 lines
Fix ao2_iterator API to hold references to containers being iterated.
See Mantis issue for details of what prompted this change.
Additional notes:
This patch changes the ao2_iterator API in two ways: F_AO2I_DONTLOCK
has become an enum instead of a macro, with a name that fits our
naming policy; also, it is now necessary to call
ao2_iterator_destroy() on any iterator that has been
created. Currently this only releases the reference to the container
being iterated, but in the future this could also release other
resources used by the iterator, if the iterator implementation changes
to use additional resources.
(closes issue #15987)
Reported by: kpfleming
Review: https://reviewboard.asterisk.org/r/383/
........
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@222176 65c4cc65-6c06-0410-ace0-fbb531ad65f3
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.
(issue #15109)
Reported by: jvandal
(issue #15123)
Reported by: axisinternet
(issue #15195)
Reported by: amorsen
(issue AST-208)
........
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@201610 65c4cc65-6c06-0410-ace0-fbb531ad65f3
Move MusicOnHold, SetMusicOnHold, StartMusicOnHold, StopMusicOnHold static
documentation to the new AstXML form.
(issue #15245)
Reported by: eliel
Patches:
res_musiconhold_static_conversion.txt uploaded by lmadsen (license 10)
(with some fixes and formatting by me)
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@199413 65c4cc65-6c06-0410-ace0-fbb531ad65f3
This patch adds 'const' tags to a number of Asterisk APIs where they are appropriate (where the API already demanded that the function argument not be modified, but the compiler was not informed of that fact). The list includes:
- CLI command handlers
- CLI command handler arguments
- AGI command handlers
- AGI command handler arguments
- Dialplan application handler arguments
- Speech engine API function arguments
In addition, various file-scope and function-scope constant arrays got 'const' and/or 'static' qualifiers where they were missing.
Review: https://reviewboard.asterisk.org/r/251/
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@196072 65c4cc65-6c06-0410-ace0-fbb531ad65f3
The moh_register function links an mohclass and then immediately
unrefs the class since the container now has a reference. The problem
with using realtime music on hold is that the class is allocated,
registered, and started in one fell swoop. The refcounting logic
resulted in the count being off by one. The same problem did not
happen when using a static config because the allocation and registration
of an mohclass is a separate operation from starting moh. This also did
not affect non-cached realtime moh because the classes are not registered
at all.
I also have modified res_musiconhold to use the _t_ variants of the ao2_
functions so that more info can be gleaned when attempting to trace the
refcounts. I found this to be incredibly helpful for debugging this issue
and there's no good reason to remove it.
(closes issue #14661)
Reported by: sum
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@187421 65c4cc65-6c06-0410-ace0-fbb531ad65f3
https://origsvn.digium.com/svn/asterisk/branches/1.4
........
r187045 | mmichelson | 2009-04-08 11:52:03 -0500 (Wed, 08 Apr 2009) | 10 lines
Fix a small logical error when loading moh classes.
We were unconditionally incrementing the number of mohclasses
registered. However, we should actually only increment if the
call to moh_register was successful.
While this probably has never caused problems, I noticed it
and decided to fix it anyway.
........
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@187046 65c4cc65-6c06-0410-ace0-fbb531ad65f3
https://origsvn.digium.com/svn/asterisk/branches/1.4
........
r181659 | file | 2009-03-12 13:50:37 -0300 (Thu, 12 Mar 2009) | 8 lines
Fix another scenario where depending on configuration the stream would not get read.
For custom commands we don't know whether the audio is coming from a stream or not
so we are going to have to read the data despite no channels.
(closes issue #14416)
Reported by: caspy
........
r181660 | file | 2009-03-12 13:52:45 -0300 (Thu, 12 Mar 2009) | 2 lines
Fix logic flaw in previous commit.
........
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@181661 65c4cc65-6c06-0410-ace0-fbb531ad65f3
https://origsvn.digium.com/svn/asterisk/branches/1.4
........
r181655 | file | 2009-03-12 13:29:19 -0300 (Thu, 12 Mar 2009) | 10 lines
Fix issue with streaming MOH failing if nobody is listening.
When a music class is setup to actually provide music on hold
from a stream we need to constantly read audio from it since it
will constantly be providing audio. This is now done despite there
being no channels listening to it.
(closes issue #14416)
Reported by: caspy
........
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@181656 65c4cc65-6c06-0410-ace0-fbb531ad65f3
https://origsvn.digium.com/svn/asterisk/branches/1.4
........
r174218 | file | 2009-02-09 10:48:21 -0400 (Mon, 09 Feb 2009) | 4 lines
Don't overwrite our pointer to the music class when music on hold stops. We will use this if it starts again to see if we can resume the music where it left off.
(closes issue #14407)
Reported by: mostyn
........
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@174219 65c4cc65-6c06-0410-ace0-fbb531ad65f3
https://origsvn.digium.com/svn/asterisk/branches/1.4
........
r165661 | russell | 2008-12-18 12:52:18 -0600 (Thu, 18 Dec 2008) | 7 lines
Set the process group ID on the MOH process so that all children will get killed
(closes issue #14099)
Reported by: caspy
Patches:
res_musiconhold.c.patch.killpg.try2 uploaded by caspy (license 645)
........
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@165662 65c4cc65-6c06-0410-ace0-fbb531ad65f3
https://origsvn.digium.com/svn/asterisk/branches/1.4
........
r162874 | jpeeler | 2008-12-10 16:04:18 -0600 (Wed, 10 Dec 2008) | 5 lines
(closes issue #13229)
Reported by: clegall_proformatique
Ensure that moh_generate does not return prematurely before local_ast_moh_stop is called. Also, the sleep in mp3_spawn now only occurs for http locations since it seems to have been added originally only for failing media streams.
........
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@162891 65c4cc65-6c06-0410-ace0-fbb531ad65f3
is unloaded and then Asterisk is stopped. The problem was that
we are not unregistering the ast_moh_destroy function at exit.
(closes issue #13761)
Reported by: eliel
Patches:
res_musiconhold.c.patch uploaded by eliel (license 64)
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@157592 65c4cc65-6c06-0410-ace0-fbb531ad65f3
when a file is invalid from when a file is missing. This is most important when
we have two configuration files. Consider the following example:
Old system:
sip.conf users.conf Old result New result
======== ========== ========== ==========
Missing Missing SIP doesn't load SIP doesn't load
Missing OK SIP doesn't load SIP doesn't load
Missing Invalid SIP doesn't load SIP doesn't load
OK Missing SIP loads SIP loads
OK OK SIP loads SIP loads
OK Invalid SIP loads incompletely SIP doesn't load
Invalid Missing SIP doesn't load SIP doesn't load
Invalid OK SIP doesn't load SIP doesn't load
Invalid Invalid SIP doesn't load SIP doesn't load
So in the case when users.conf doesn't load because there's a typo that
disrupts the syntax, we may only partially load users, instead of failing with
an error, which may cause some calls not to get processed. Worse yet, the old
system would do this with no indication that anything was even wrong.
(closes issue #10690)
Reported by: dtyoo
Patches:
20080716__bug10690.diff.txt uploaded by Corydon76 (license 14)
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@142992 65c4cc65-6c06-0410-ace0-fbb531ad65f3
https://origsvn.digium.com/svn/asterisk/branches/1.4
........
r125132 | kpfleming | 2008-06-25 17:21:30 -0500 (Wed, 25 Jun 2008) | 10 lines
allow tonezone to live in a different place than DAHDI/Zaptel, since dahdi-tools and dahdi-linux are now separate packages and can be installed in different places
don't include tonezone.h in dahdi_compat.h, because only a couple of modules need it
get app_rpt building again after the DAHDI changes
(closes issue #12911)
Reported by: tzafrir
........
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@125138 65c4cc65-6c06-0410-ace0-fbb531ad65f3
They want (char *)NULL as sentinel.
An example is OpenBSD (confirmed on 4.3) that ships with gcc 3.3.4
This commit introduces a contstant SENTINEL which is declared as:
#define SENTINEL ((char *)NULL)
All places I could test compile on my openbsd system are converted.
Update CODING-GUIDELINES to tell about this constant.
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@124127 65c4cc65-6c06-0410-ace0-fbb531ad65f3
- make data member of the ast_frame struct a named union instead of a void
Recently the ast_queue_hangup function got a new parameter, the hangupcause
Feedback came in that this is no good and that instead a new function should be created.
This I did.
The hangupcause was stored in the seqno member of the ast_frame struct. This is not very
elegant, and since there's already a data member that one should be used.
Problem is, this member was a void *.
Now it's a named union so it can hold a pointer, an uint32 and there's a padding in case someone
wants to store another type in there in the future.
This commit is so massive, because all ast_frame.data uses have to be
altered to ast_frame.data.data
Thanks russellb and kpfleming for the feedback.
(closes issue #12674)
Reported by: mvanbaak
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@117802 65c4cc65-6c06-0410-ace0-fbb531ad65f3
https://origsvn.digium.com/svn/asterisk/branches/1.4
........
r115418 | qwell | 2008-05-06 14:34:58 -0500 (Tue, 06 May 2008) | 7 lines
Switch to using ast_random() rather than just rand().
This does not fix the bug reported, but I believe it is correct.
(from issue #12446)
Patches:
bug_12446.diff uploaded by snuffy (license 35)
........
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@115419 65c4cc65-6c06-0410-ace0-fbb531ad65f3
G.722 music on hold working for me.
(issue #12164, reported by milazzo and jsmith, patches by me)
res/res_musiconhold.c:
- I moved a single line so that the sample queue update happened before
ast_write(). The reason that this was a bug is that the G.722 frame
originally says it has 320 samples in it (which is correct). However,
when the frame is written to a channel that uses RTP, main/rtp.c modifies
the frame to cut the number of samples in half before it sends it on
the wire. This is to account for the stupid incorrect G.722 spec that
makes it so we have to lie about the number of samples with RTP. I should
probably go and re-work the RTP code so it doesn't modify the frame so
that a bug like this won't happen in the future. However, this change to
MOH is harmless.
main/channel.c:
- I made two fixes in regards to generator timing. Generators use samples
for timing. However, this code assumed 8 kHz samples. In one case, it was
a hard coded 160 samples, that is now written as the sample rate / 50. The
other place was dealing with timing a generator based on frames coming from
the other direction. However, that would have only worked if the sample
rates for the formats in both directions were the same. The code now takes
into account that the sample rates may differ, and scales the generator
samples accordingly.
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@110268 65c4cc65-6c06-0410-ace0-fbb531ad65f3
https://origsvn.digium.com/svn/asterisk/branches/1.4
........
r110035 | file | 2008-03-19 16:11:33 -0300 (Wed, 19 Mar 2008) | 4 lines
Add sanity checking for position resuming. We *have* to make sure that the position does not exceed the total number of files present, and we have to make sure that the position's filename is the same as previous. These values can change if a music class is reloaded and give unpredictable behavior.
(closes issue #11663)
Reported by: junky
........
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@110036 65c4cc65-6c06-0410-ace0-fbb531ad65f3
1. When moh is started, we search first in memory to find the class. If we do not
find it in memory, we search realtime instead.
2. When moh is restarted (as in, it had been started on this particular channel, stopped,
and now we're starting it again), if using the "files" mode, then realtime will always
be rechecked. If you are using other modes, however, we will simply reattach to the external
running process which was playing moh earlier in the call. This is a necessary compromise so that
we don't end up with too many background processes.
3. musiconhold.conf has a general section now. It has one option: cachertclasses. If set to yes,
then moh classes found in realtime will be added to the in-memory list. This has the advantage
of not requiring database lookups each time moh is started, but it has the disadvantage of not
truly being realtime.
I have tested this for functionality, and it passes. I also tested this under valgrind and there
are no memory problems reported under typical use.
Special thanks to Sergee for implementing this feature and enduring my complaints on the bugtracker!
(closes issue #11196, reported and patched by sergee)
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@89946 65c4cc65-6c06-0410-ace0-fbb531ad65f3
build times - tested, there is no measureable difference before and
after this commit.
In this change:
use asterisk/compat.h to include a small set of system headers:
inttypes.h, unistd.h, stddef.h, stddint.h, sys/types.h, stdarg.h,
stdlib.h, alloca.h, stdio.h
Where available, the inclusion is conditional on HAVE_FOO_H as determined
by autoconf.
Normally, source files should not include any of the above system headers,
and instead use either "asterisk.h" or "asterisk/compat.h" which does it
better.
For the time being I have left alone second-level directories
(main/db1-ast, etc.).
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@89333 65c4cc65-6c06-0410-ace0-fbb531ad65f3
- the *_CURRENT macros no longer need the list head pointer argument
- add AST_LIST_MOVE_CURRENT to encapsulate the remove/add operation when moving entries between lists
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@89106 65c4cc65-6c06-0410-ace0-fbb531ad65f3
https://origsvn.digium.com/svn/asterisk/branches/1.4
........
r89037 | russell | 2007-11-06 12:20:07 -0600 (Tue, 06 Nov 2007) | 11 lines
If someone were to delete the files used by an existing MOH class, and then
issue a reload, further use of that class could result in a crash due to
dividing by zero. This set of changes fixes up some places to prevent this
from happening.
(closes issue #10948)
Reported by: jcomellas
Patches:
res_musiconhold_division_by_zero.patch uploaded by jcomellas (license 282)
Additional changes added by me.
........
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@89038 65c4cc65-6c06-0410-ace0-fbb531ad65f3
https://origsvn.digium.com/svn/asterisk/branches/1.4
........
r84160 | file | 2007-10-01 10:57:42 -0300 (Mon, 01 Oct 2007) | 6 lines
Fix randomness. save_pos was being set to 0 initially instead of -1, causing it to jump to position 0 when moh started.
(closes issue #10859)
Reported by: jamesgolovich
Patches:
asterisk-mohpos2.diff.txt uploaded by jamesgolovich (license 176)
........
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@84161 65c4cc65-6c06-0410-ace0-fbb531ad65f3
https://origsvn.digium.com/svn/asterisk/branches/1.4
........
r81042 | russell | 2007-08-27 11:16:25 -0500 (Mon, 27 Aug 2007) | 11 lines
(closes issue #10419)
Reported by: mustardman
Patches:
asterisk-mohposition.diff.txt uploaded by jamesgolovich (license 176)
This patch fixes a few problems with music on hold.
* Fix issues with starting at the beginning of a file when it shouldn't.
* Fix the inuse counter to be decremented even if the class had not been
set to be deleted when not in use anymore
* Don't arbitrarily limit the number of MOH files to 255
........
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@81043 65c4cc65-6c06-0410-ace0-fbb531ad65f3
https://origsvn.digium.com/svn/asterisk/branches/1.4
........
r79778 | russell | 2007-08-16 17:24:25 -0500 (Thu, 16 Aug 2007) | 14 lines
This patch fixes a bug where reloading the module with "module reload" did not
delete classes from memory that were no longer in the config. This patch fixes
that problem as well as another one. Previously, if you reloaded MOH using the
"moh reload" CLI command, which behaved differently than "module reload ...",
MOH had to be stopped on every channel and started again immediately. However,
there was no way to tell what class was being used, so they would all fall back
to the default class.
(closes issue #10139)
Reported by: blitzrage
Patches:
asterisk-10139-advanced.diff.txt uploaded by jamesgolovich (license 176)
Tested by: jamesgolovich
........
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@79788 65c4cc65-6c06-0410-ace0-fbb531ad65f3
https://origsvn.digium.com/svn/asterisk/branches/1.4
................
r75067 | russell | 2007-07-13 15:10:40 -0500 (Fri, 13 Jul 2007) | 14 lines
Merged revisions 75059 via svnmerge from
https://origsvn.digium.com/svn/asterisk/branches/1.2
........
r75059 | russell | 2007-07-13 15:07:21 -0500 (Fri, 13 Jul 2007) | 6 lines
Ensure that adding a user to the list of users of a specific music on hold
class is not done at the same time as any of the other operations on this list
to prevent list corruption. Using the global moh_data lock for this is not
ideal, but it is what is used to protect these lists everywhere else in the
module, and I am only changing what is necessary to fix the bug.
........
................
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@75075 65c4cc65-6c06-0410-ace0-fbb531ad65f3
https://origsvn.digium.com/svn/asterisk/branches/1.4
........
r74162 | russell | 2007-07-09 15:53:46 -0500 (Mon, 09 Jul 2007) | 9 lines
(closes issue #10123)
Reported by: blitzrage
Patches submitted by: juggie, qwell, me
Tested by: blitzrage
When trying to find a music on hold class to use, try all of the options,
instead of only the first one that is set. Also, change the MusicOnHold
applications to not hang up on the channel when a class can not be found.
........
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@74163 65c4cc65-6c06-0410-ace0-fbb531ad65f3
class in musiconhold.conf, a caller on hold may press this digit to switch
to listening to that music class.
This involved adding a new callback for generators, which allow generators
to get notified of DTMF from the channel they are running on. Then, a callback
was implemented for the music on hold generators.
(patch from bbryant)
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@65505 65c4cc65-6c06-0410-ace0-fbb531ad65f3
https://origsvn.digium.com/svn/asterisk/branches/1.4
................
r48375 | tilghman | 2006-12-10 18:47:21 -0600 (Sun, 10 Dec 2006) | 13 lines
Merged revisions 48374 via svnmerge from
https://origsvn.digium.com/svn/asterisk/branches/1.2
........
r48374 | tilghman | 2006-12-10 18:33:59 -0600 (Sun, 10 Dec 2006) | 5 lines
When doing a fork() and exec(), two problems existed (Issue 8086):
1) Ignored signals stayed ignored after the exec().
2) Signals could possibly fire between the fork() and exec(), causing Asterisk
signal handlers within the child to execute, which caused nasty race conditions.
........
................
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@48376 65c4cc65-6c06-0410-ace0-fbb531ad65f3
https://origsvn.digium.com/svn/asterisk/branches/1.4
................
r47239 | russell | 2006-11-06 20:25:10 -0500 (Mon, 06 Nov 2006) | 13 lines
Merged revisions 47238 via svnmerge from
https://origsvn.digium.com/svn/asterisk/branches/1.2
........
r47238 | russell | 2006-11-06 20:22:58 -0500 (Mon, 06 Nov 2006) | 5 lines
If random order is enabled for files mode music on hold, set a random initial
position, instead of always starting at the first file, and doing the random
operation only when switching to the next file.
(bug reported by John Lange on the asterisk-dev mailing list)
........
................
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@47240 65c4cc65-6c06-0410-ace0-fbb531ad65f3
https://origsvn.digium.com/svn/asterisk/branches/1.4
........
r46363 | russell | 2006-10-27 12:39:31 -0500 (Fri, 27 Oct 2006) | 5 lines
We should always be using _exit() after a fork() or vfork() instead of exit().
This is because exit() does some extra cleanup which in some implementations
of vfork(), for example, can actually modify the state of the parent process,
causing very weird bugs or crashes. (issue #7971, Nick Gavrikov)
........
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@46364 65c4cc65-6c06-0410-ace0-fbb531ad65f3
https://origsvn.digium.com/svn/asterisk/branches/1.4
........
r44378 | kpfleming | 2006-10-04 14:47:22 -0500 (Wed, 04 Oct 2006) | 4 lines
update thread creation code a bit
reduce standard thread stack size slightly to allow the pthreads library to allocate the stack+data and not overflow a power-of-2 allocation in the kernel and waste memory/address space
add a new stack size for 'background' threads (those that don't handle PBX calls) when LOW_MEMORY is defined
........
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@44379 65c4cc65-6c06-0410-ace0-fbb531ad65f3
- restructured build tree and makefiles to eliminate recursion problems
- support for embedded modules
- support for static builds
- simpler cross-compilation support
- simpler module/loader interface (no exported symbols)
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@40722 65c4cc65-6c06-0410-ace0-fbb531ad65f3
Tilghman Lesher
Kevin P. Fleming
Matthias Nick
Mark Michelson
Russell Bryant
Eliel C. Sardanons
Joshua Colp
Jeff Peeler
Sean Bright
Michiel van Baak
Jason Parker
Luigi Rizzo
Steve Murphy
Matt O'Gorman
Matthew Fredrickson