Joshua Colp [Mon, 6 Aug 2007 14:37:52 +0000 (14:37 +0000)]
(closes issue #10381)
Reported by: yehavi
Use the filename we parsed using the standard parsing when launching the application specified to ExternalIVR.
Russell Bryant [Sun, 5 Aug 2007 03:14:24 +0000 (03:14 +0000)]
Fix building res_crypto on systems that init locks with constructors.
The problem was that res_crypto now has a RWLIST named "keys". The macro
for defining this list defines a function used as a constructor for the list
called "init_keys". However, there was another function called init_keys in
this module for a CLI command. The fix is just to prepend the generated
functions with underscores.
Fix a problem when you call Voicemail() with multiple mailboxes specified and
ODBC_STORAGE is in use. The audio part of the message was only given to the
first mailbox specified.
Russell Bryant [Fri, 3 Aug 2007 19:41:42 +0000 (19:41 +0000)]
Merged revisions 78095 via svnmerge from
https://origsvn.digium.com/svn/asterisk/branches/1.4
........
r78095 | russell | 2007-08-03 14:39:49 -0500 (Fri, 03 Aug 2007) | 28 lines
Add some improvements to lock debugging. These changes take effect
with DEBUG_THREADS enabled and provide the following:
* This will keep track of which locks are held by which thread as well as
which lock a thread is waiting for in a thread-local data structure. A
reference to this structure is available on the stack in the dummy_start()
function, which is the common entry point for all threads. This information
can be easily retrieved using gdb if you switch to the dummy_start() stack
frame of any thread and print the contents of the lock_info variable.
* All of the thread-local structures for keeping track of this lock information
are also stored in a list so that the information can be dumped to the CLI
using the "core show locks" CLI command. This introduces a little bit of a
performance hit as it requires additional underlying locking operations
inside of every lock/unlock on an ast_mutex. However, the benefits of
having this information available at the CLI is huge, especially considering
this is only done in DEBUG_THREADS mode. It means that in most cases where
we debug deadlocks, we no longer have to request access to the machine to
analyze the contents of ast_mutex_t structures. We can now just ask them
to get the output of "core show locks", which gives us all of the information
we needed in most cases.
I also had to make some additional changes to astmm.c to make this work when
both MALLOC_DEBUG and DEBUG_THREADS are enabled. I disabled tracking of one
of the locks in astmm.c because it gets used inside the replacement memory
allocation routines, and the lock tracking code allocates memory. This caused
infinite recursion.
Russell Bryant [Fri, 3 Aug 2007 17:02:00 +0000 (17:02 +0000)]
Merged revisions 78063 via svnmerge from
https://origsvn.digium.com/svn/asterisk/branches/1.4
........
r78063 | russell | 2007-08-03 12:01:07 -0500 (Fri, 03 Aug 2007) | 4 lines
Only pass through HOLD and UNHOLD control frames when the mohinterpret option
is set to "passthrough". This was pointed out by Kevin in the middle of a
training session.
Russell Bryant [Fri, 3 Aug 2007 02:05:02 +0000 (02:05 +0000)]
Merged revisions 78028 via svnmerge from
https://origsvn.digium.com/svn/asterisk/branches/1.4
........
r78028 | russell | 2007-08-02 21:04:22 -0500 (Thu, 02 Aug 2007) | 6 lines
Don't reuse the timespec that was set to 0 in the previous timedwait as it
will just return immediately. Also, fix some logic so the thread's lock
isn't unlocked twice in the weird case of dynamic threads getting acquired
right after a timeout.
(pointed out by SteveK)
Russell Bryant [Thu, 2 Aug 2007 19:26:12 +0000 (19:26 +0000)]
Merged revisions 77949 via svnmerge from
https://origsvn.digium.com/svn/asterisk/branches/1.4
........
r77949 | russell | 2007-08-02 14:25:14 -0500 (Thu, 02 Aug 2007) | 5 lines
Fix the case where a dynamic thread times out waiting for something to do
during the first time it runs. This shouldn't ever happen, but we should
account for it anyway.
(pointed out by pete, who works with mihai)
Merged revisions 77942 via svnmerge from
https://origsvn.digium.com/svn/asterisk/branches/1.2
........
r77942 | murf | 2007-08-02 11:56:37 -0600 (Thu, 02 Aug 2007) | 1 line
This patch hopefully solves 10141; The user is running with it, and it doesn't appear to harm asterisk's operation, and may prevent a crash. I'll store it in 1.2, as we have shut down support on 1.2, but since I developed the patch before support finished, and it might affect 1.4 and trunk, I'm going ahead with it.
........
Russell Bryant [Thu, 2 Aug 2007 18:05:23 +0000 (18:05 +0000)]
Merged revisions 77943 via svnmerge from
https://origsvn.digium.com/svn/asterisk/branches/1.4
........
r77943 | russell | 2007-08-02 13:04:43 -0500 (Thu, 02 Aug 2007) | 9 lines
Fix another race condition in the handling of dynamic threads. If the dynamic
thread timed out waiting for something to do, but was acquired to perform an
action immediately afterwords, then wait on the condition again to give the
other thread a chance to finish setting up the data for what action this thread
should perform. Otherwise, if it immediately continues, it will perform the
wrong action.
(reported on IRC by mihai, patch by me)
(related to issue #10289)
Russell Bryant [Thu, 2 Aug 2007 17:09:42 +0000 (17:09 +0000)]
Fix an issue that Simon pointed out to me on IRC. There were cases in the
trunk version of find_idle_thread() where the old full frame processing
information was not cleared out. This would have caused full frames to get
deferred for processing by threads that weren't actually processing frames for
that call. Nice catch!!
Russell Bryant [Thu, 2 Aug 2007 16:58:26 +0000 (16:58 +0000)]
Merged revisions 77939 via svnmerge from
https://origsvn.digium.com/svn/asterisk/branches/1.4
........
r77939 | russell | 2007-08-02 11:56:04 -0500 (Thu, 02 Aug 2007) | 4 lines
Add another sanity check to vnak_retransmit(). This check ensures that frames
that have already been marked for deletion don't get retransmitted.
(closes issue #10361, patch from mihai)
Russell Bryant [Wed, 1 Aug 2007 22:24:23 +0000 (22:24 +0000)]
Merged revisions 77887 via svnmerge from
https://origsvn.digium.com/svn/asterisk/branches/1.4
........
r77887 | russell | 2007-08-01 17:16:17 -0500 (Wed, 01 Aug 2007) | 23 lines
Fix some race conditions which have been causing weird problems in chan_iax2.
The most notable problem is that people have been seeing storms of VNAK frames
being sent due to really old frames mysteriously being in the retransmission
queue and never getting removed.
It was possible that a dynamic thread got created, but did not acquire its lock
before the thread that created it signals it to perform an action. When this
happens, the thread will sleep until it hits a timeout, and then get destroyed.
So, the action never gets performed and in some cases, means a frame doesn't
get transmitted and never gets freed since the scheduler never gets a chance
to reschedule transmission.
Another less severe race condition is in the handling of a timeout for a dynamic
thread. It was possible for it to be acquired to perform at action at the same
time that it hit a timeout. When this occurs, whatever action it was acquired
for would never get performed.
(patch contributed by Mihai and SteveK)
(closes issue #10289)
(closes issue #10248)
(closes issue #10232)
(possibly related to issue #10359)
Joshua Colp [Wed, 1 Aug 2007 19:37:59 +0000 (19:37 +0000)]
Instead of adding the SOLARIS check to each HAVE_SYSINFO check let's just make the sysinfo autoconf logic a bit pickier about what it considers a usable sysinfo.
Russell Bryant [Wed, 1 Aug 2007 15:39:54 +0000 (15:39 +0000)]
Convert code that checks the _softhangup member of ast_channel directory to use
the ast_check_hangup() funciton. This function takes scheduled hangups into
account.
(closes issue #10230, patch by Juggie)
Fixes an issue I introduced to queues wherein a queue with joinempty=yes would kick people out of the queue because of erroneously
thinking the 'n' option was in use.
(closes issue #10320, reported by jfitzgibbon, patched by me, tested by blitzrage and me)
Thank you blitzrage for all the testing you've done lately with queues! It's much appreciated!
If a queue uses dynamic realtime members, then the member list should be updated after each attempt to call the queue.
This fixes an issue where if a caller calls into a queue where no one is logged in, they would wait forever even if a member
logged in at some point.
(closes issue #10346, reported by and tested by blitzrage, patched by me)
This patch makes Asterisk send 100 Trying provisional responses upon receipt of re-invites. This makes it so that if there are two or more Asterisk
servers between endpoints, the Asterisk servers will not keep retransmitting the re-invites.
(closes issue #10274, reported by cstadlmann, patched by me with approval from file)
Kevin P. Fleming [Tue, 31 Jul 2007 15:01:27 +0000 (15:01 +0000)]
there is no use in having functions that have no code in them, and hide the locking info when DEBUG_THREADS is enabled... i could have fixed this to be dependent on DEBUG_THREADS, but it would be just as easy for someone to add their test/debugging code to the macros as it would have been to the functions
Kevin P. Fleming [Tue, 31 Jul 2007 14:55:37 +0000 (14:55 +0000)]
use a different method for overriding the send_digit_begin pointer, as the old one fails to compile on my 64-bit system with gcc-4.1 and --enable-dev-mode turned on
Steve Murphy [Tue, 31 Jul 2007 03:32:04 +0000 (03:32 +0000)]
Discovered in experiments on core files: if you wrap the lock and unlock calls with sip_pvt_lock and sip_pvt_unlock, you lose the tracing info you would normally get via DETECT_DEADLOCKS; so I turn these two functions into macros when DETECT_DEADLOCKS is called. This way, you get meaningful stuff in the file and func slots in the lock_info struct.
Fix an issue that could potentially cause corruption of the global iax frame
queue. In the network_thread() loop, it traverses the list using the
AST_LIST_TRAVERSE_SAFE macro. However, to remove an element of the list within
this loop, it used AST_LIST_REMOVE, instead of AST_LIST_REMOVE_CURRENT, which I
believe could leave some of the internal variables of the SAFE macro invalid.
Mihai says that he already made this change in his local copy and it didn't help
his VNAK storm issues, but I still think it's wrong. :)
Applications like SayAlpha() should not hang up the channel if you
request an "unknown" character such as a comma.
Instead, skip the character and move on.
Issue 10083, initial patch by jsmith, modified by me.
Russell Bryant [Mon, 30 Jul 2007 19:35:33 +0000 (19:35 +0000)]
Improve ast_agi_fdprintf() by using the ast_str() API.
* Use a thread local ast_str for building the string that will be written out
to the console for debug, and to the FD for the AGI itself, instead of allocating
a buffer on the heap every time the function is called.
* Use the information contained within the ast_str to determine how many bytes
need to be written instead of calling strlen().
Russell Bryant [Mon, 30 Jul 2007 19:31:27 +0000 (19:31 +0000)]
Remove an XXX comment noting that it would be nice for a declaration to be
inside of a function. (Yes, it would!) Replace it with a note that explains
why it can't be done using the way that the AST_THREADSTORAGE macro is
currently defined.
(closes issue #10279)
Reported by: seanbright
Patches:
res_agi.carefulwrite.1.4.07252007.patch uploaded by seanbright (license 71)
res_agi.carefulwrite.trunk.07252007.patch uploaded by seanbright (license 71)
Allow the "agi_network: yes" line to be printed out in the AGI debug output.
Also, allow partial writes to be handled when writing out this line just like
it is for all of the others.
(closes issue #10301)
Reported by: fnordian
Patches:
asterisk-1.4.9-channel.c.patch uploaded by fnordian (license 110)
Additional changes by me
Fix some problems in channel_find_locked() which can cause an infinite loop.
The reference to the previous channel is set to NULL in some cases. These changes
ensure that the reference to the previous channel gets restored before needing
it again.
I'm not convinced that the code that is setting it to NULL is really the right
thing to do. However, I am making these changes to fix the obvious problem
and just leaving an XXX comment that it needs a better explanation that what
is there now.
(closes issue #10327)
Reported by: kkiely
Instead of directly mucking with the extension/context/priority of the channel we are transferring when it has a PBX simply call ast_async_goto on it. This will ensure that the channel gets handled properly and sent to the right place.
(closes issue #10301)
Reported by: fnordian
Patches:
asterisk-1.4.9-channel.c.patch uploaded by fnordian (license 110)
Restore previous behavior where if we failed to lock the channel we wanted we would return to exactly the same point as if we had just reentered the function.
Luigi Rizzo [Mon, 30 Jul 2007 10:55:37 +0000 (10:55 +0000)]
minor code rearrangements:
+ place the link field at the beginning of struct sip_pvt,
and not somewhere in the middle;
+ in __sip_reliable_xmit, remove a duplicate assignment, and
put the statements in a more logical order (i.e. first copy
the payload and associated info, then copy arguments from the
caller, then finish initializing the headers...)
Luigi Rizzo [Mon, 30 Jul 2007 08:07:00 +0000 (08:07 +0000)]
rename handle_request() to handle_incoming(), as the former
was misleading - the function deals with all incoming packets, be
them requests or responses.
Luigi Rizzo [Sun, 29 Jul 2007 21:24:56 +0000 (21:24 +0000)]
move some dialog-only flags to proper variables, namely
SIP_NOVIDEO, SIP_DIALOG_ANSWEREDELSEWHERE, SIP_PAGE2_NOTEXT,
SIP_PAGE2_OUTGOING_CALL
These are seldom used so the diff is relatively small.
Note that 'OUTGOING_CALL' is dangerously similar to another
dialog flag, 'SIP_OUTGOING', so the description will need to
clarify the different meaning of the two.
Also note that the description of NOTEXT is a bit unclear - does
it mean we don't support it, or 'not requested or not supported' ?
On passing fix a comment referring to video instead of text.
Finally, mark with XXX a possibly misleading debugging message.
(maybe the latter is worth backporting).
Luigi Rizzo [Sun, 29 Jul 2007 20:55:20 +0000 (20:55 +0000)]
use a function, cli_yesno(), to produce the output Yes or No for
CLI lines. This helps maintaining consistency on output, slightly
improves readability, and maybe one day will make it easier to
translate the output in other languages (though i have a hard time
believing that a CLI user who needs 'yes' and 'no' to be translated
can actually figure out what he/she is doing!)
Luigi Rizzo [Sun, 29 Jul 2007 20:01:36 +0000 (20:01 +0000)]
Move some global 'flags' to individual variables.
Start putting these variables in a single struct (called 'sip_cfg' for the time
being, but it could as well be 'global' or some other name) so it
is easy, when reading the code, to figure out what they are for.
The downside of using struct fields instead of individual global
variables is that the compiler cannot tell if there are unused fields.
But the advantage of not cluttering the namespace and manilpulating
all these variables at once certainly overcome the disadvantagess.
Luigi Rizzo [Sun, 29 Jul 2007 10:13:14 +0000 (10:13 +0000)]
build the version of sip_tech with no send_digit_begin
at load time instead of duplicating the initializer.
This should remove the risk of forgetting fields in the
initializer.
Luigi Rizzo [Sun, 29 Jul 2007 09:27:30 +0000 (09:27 +0000)]
remove bit position from description of SIP_* flags.
use AST_FORMAT_AUDIO_MASK instead of playing with AST_FORMAT_MAX_AUDIO
to determine audio formats.
There is a dubious use of AST_FORMAT_MAX_AUDIO in sip_request_call()
which surely needs fixing, namely:
/* mask request with some set of allowed formats.
* XXX this needs to be fixed.
* The original code uses AST_FORMAT_AUDIO_MASK, but it is
* unclear what to use here. We have global_capabilities, which is
* configured from sip.conf, and sip_tech.capabilities, which is
* hardwired to all audio formats.
*/
The latter is possibly something to backport when fixed.
Luigi Rizzo [Sun, 29 Jul 2007 08:58:10 +0000 (08:58 +0000)]
back on cleaning up the usage of flags.
Move together flags used in the same way (e.g. dialog only,
dialog-peer, ...) so it will become easier to deal with them
in a more systematic way.
This is being done in stages so it will be easier to detect
breakage, if any should occur.
Luigi Rizzo [Sat, 28 Jul 2007 23:43:35 +0000 (23:43 +0000)]
add some documentation to auto_congest(), and some
dialog_ref/unref (they are a no-op at the moment).
Also clean a pointer after freeing memory to avoid
dangling references, and write a for() loop in canonical form.
In practice, everything in this commit is a no-op.
Luigi Rizzo [Sat, 28 Jul 2007 16:25:25 +0000 (16:25 +0000)]
start introducing hooks for reference counts on dialog descriptors.
This commit is, for all practical purposes, a no-op, as it only
introduces the dialog_ref() and dialog_unref() methods, and uses them
in a few places (not all the places where they would be needed).
The goal is to start annotating the code with these calls, so the transition
to a proper container will be easier.
Luigi Rizzo [Sat, 28 Jul 2007 07:44:16 +0000 (07:44 +0000)]
make use of received= and rport= fields in sip replies.
In a nutshell, these fields are used to tell a sip entity
the address and port its request came from, and are extremely
useful in the presence of NATs, especially with symmetric NATs
where STUN is totally ineffective.
This patch stores the address and port in the 'ourip' field of
the dialog descriptor, so they can be reused in subsequent transactions.
As it is, it works well for things like REGISTER requiring authentication,
because the second REGISTER request (with auth credentials) will carry
the correct address. Maybe it can also be useful, in case of an address
change, to do one or both of the following:
+ propagate the new address to the parent user/peer descriptor so that new
dialogs will use the correct address from the beginning.
This is trivial to implement, I am just waiting for feedback on this.
+ re-issue a request in case of an address change. This a lot less trivial,
maybe unnecessary, and probably covered by the previous item.
I would seriously consider this patch for addition to 1.4 and 1.2.
The code is very little intrusive, and it would solve in a correct
way the nat traversal problems for which externip/externaddr/stunaddr
are only a partial and expensive workaround.
(closes issue #10310)
Reported by: prashant_jois
Patches:
cdr_pgsql.patch uploaded by prashant (license 114)
Finish the Postgresql connection after the log messages are printed so we don't access invalid memory.
(closes issue #10323)
Reported by: julianjm
Patches:
chan_sip_device_state_hold_fix.v1.diff.txt uploaded by julianjm (license 99)
Clear ONHOLD flag when decrementing the onHold peer count. If we did not do this the count may keep decreasing.
Steve Murphy [Fri, 27 Jul 2007 15:46:20 +0000 (15:46 +0000)]
These fixes take care of two problems: a complaint in asterisk-dev that goto's aren't working in trunk, a side effect of the move to commas as arg seps in apps and funcs; and a problem I spotted myself with dial's 'e' option, where gotos were off by one, because I forgot to set the AUTOLOOP flag in the peer channel.
(closes issue #10302)
Reported by: litnialex
If a DTMF end frame comes from a channel without a begin and it is going to a technology that only accepts end frames (aka INFO) then use the minimum DTMF duration if one is not in the frame already.
projects / thirdparty / asterisk.git / log
