]> git.ipfire.org Git - thirdparty/freeswitch.git/log
thirdparty/freeswitch.git
11 years agorebuild
Anthony Minessale [Fri, 11 Jul 2014 23:39:56 +0000 (04:39 +0500)] 
rebuild

11 years agomerge ws.c change to sofia
Anthony Minessale [Fri, 11 Jul 2014 23:39:41 +0000 (04:39 +0500)] 
merge ws.c change to sofia

11 years agocss tweak
Anthony Minessale [Fri, 11 Jul 2014 23:29:15 +0000 (04:29 +0500)] 
css tweak

11 years agotweak on last commit
Anthony Minessale [Fri, 11 Jul 2014 23:17:41 +0000 (04:17 +0500)] 
tweak on last commit

11 years agoadd clear channel for mono participants while in a conference with 3d callers
Anthony Minessale [Fri, 11 Jul 2014 22:31:26 +0000 (03:31 +0500)] 
add clear channel for mono participants while in a conference with 3d callers

11 years agofold livetables into verto obj and expose moderator mode
Anthony Minessale [Fri, 11 Jul 2014 22:27:03 +0000 (03:27 +0500)] 
fold livetables into verto obj and expose moderator mode

11 years agoFS-6642 --resolve
Brian West [Fri, 11 Jul 2014 16:04:08 +0000 (11:04 -0500)] 
FS-6642 --resolve

11 years agoFS-6646 --resolve
Brian West [Fri, 11 Jul 2014 16:03:22 +0000 (11:03 -0500)] 
FS-6646 --resolve

11 years agomod_graylog2: additional input checks on LOG_FIELDS[]
Chris Rienzo [Fri, 11 Jul 2014 14:38:29 +0000 (10:38 -0400)] 
mod_graylog2: additional input checks on LOG_FIELDS[]

11 years agofreetdm: ftmod_gsm: Added support for hardware DTMF
Moises Silva [Fri, 11 Jul 2014 04:41:40 +0000 (00:41 -0400)] 
freetdm: ftmod_gsm: Added support for hardware DTMF

11 years agoFS-6420
Brian West [Thu, 10 Jul 2014 15:17:38 +0000 (10:17 -0500)] 
FS-6420

11 years agomod_graylog2: additional fields can now be specified by adding LOG_FIELDS[name=string...
Chris Rienzo [Thu, 10 Jul 2014 14:25:20 +0000 (10:25 -0400)] 
mod_graylog2: additional fields can now be specified by adding LOG_FIELDS[name=string_value,@#name=number_value] to beginning of log.  Added _microtimestamp additional field since graylog2 only has millisecond timestamp precision.

11 years agolog websocket write failures
Anthony Minessale [Wed, 9 Jul 2014 18:30:04 +0000 (13:30 -0500)] 
log websocket write failures

11 years agoCID:1224272 Dereference after null check. member can not actually be NULL in any...
Michael Jerris [Wed, 9 Jul 2014 18:32:38 +0000 (14:32 -0400)] 
CID:1224272 Dereference after null check.  member can not actually be NULL in any of the scenarios calling this function

11 years agoMerge pull request #9 in FS/freeswitch from ~PFOURNIER/freeswitch:feature/fax_enable_...
Mike Jerris [Wed, 9 Jul 2014 16:54:16 +0000 (11:54 -0500)] 
Merge pull request #9 in FS/freeswitch from ~PFOURNIER/freeswitch:feature/fax_enable_tep to master

* commit '0fb179acf2246b896f1a66cfb96da68e9d2675f7':
  Add new option to request TEP to be used for an outbound fax call

11 years agoMerge pull request #7 in FS/freeswitch from ~PFOURNIER/freeswitch:bugfix/t38_require_...
Brian West [Wed, 9 Jul 2014 15:49:04 +0000 (10:49 -0500)] 
Merge pull request #7 in FS/freeswitch from ~PFOURNIER/freeswitch:bugfix/t38_require_timer to master

* commit '21ae587063eb3b55c690f47cf9299a539ec25415':
  Disabling Require timer for T.38 re-Invites cause problems

11 years agoMerge pull request #8 in FS/freeswitch from ~PFOURNIER/freeswitch:bugfix/fax_disable_...
Brian West [Wed, 9 Jul 2014 15:48:56 +0000 (10:48 -0500)] 
Merge pull request #8 in FS/freeswitch from ~PFOURNIER/freeswitch:bugfix/fax_disable_ecm to master

* commit '5d31c9b9fe2ae6dead163c60974a41fec4fac3fc':
  It was impossible to disable ECM with T.38 in gateway mode

11 years agoFS-4861 improve logging when using c++ wrapper
Brian West [Wed, 9 Jul 2014 14:09:45 +0000 (09:09 -0500)] 
FS-4861 improve logging when using c++ wrapper

11 years agoFS-4870
Brian West [Wed, 9 Jul 2014 14:08:05 +0000 (09:08 -0500)] 
FS-4870

11 years agoFS-4939
Brian West [Wed, 9 Jul 2014 14:04:17 +0000 (09:04 -0500)] 
FS-4939

11 years agoFS-5501 fire event on profile start
Brian West [Wed, 9 Jul 2014 13:53:14 +0000 (08:53 -0500)] 
FS-5501 fire event on profile start

11 years agoFS-5516 allow get params in URL
Brian West [Wed, 9 Jul 2014 13:52:01 +0000 (08:52 -0500)] 
FS-5516 allow get params in URL

11 years agoFS-6373
Brian West [Wed, 9 Jul 2014 13:25:54 +0000 (08:25 -0500)] 
FS-6373

11 years agoFS-6626 fix winter DST issue
Brian West [Wed, 9 Jul 2014 12:58:11 +0000 (07:58 -0500)] 
FS-6626 fix winter DST issue

11 years agoFS-6563
Brian West [Wed, 9 Jul 2014 12:47:17 +0000 (07:47 -0500)] 
FS-6563

11 years agoFS-6565
Brian West [Wed, 9 Jul 2014 12:45:55 +0000 (07:45 -0500)] 
FS-6565

11 years agoFS-6566
Brian West [Wed, 9 Jul 2014 12:44:33 +0000 (07:44 -0500)] 
FS-6566

11 years agoAdded explanatory comment on a recent change.
Steve Underwood [Wed, 9 Jul 2014 02:18:00 +0000 (10:18 +0800)] 
Added explanatory comment on a recent change.

11 years agoMerge branch 'master' of ssh://stash.freeswitch.org:7999/fs/freeswitch
Steve Underwood [Tue, 8 Jul 2014 16:14:31 +0000 (00:14 +0800)] 
Merge branch 'master' of ssh://stash.freeswitch.org:7999/fs/freeswitch

11 years agosync changes from srtp upstream
Michael Jerris [Tue, 8 Jul 2014 14:25:52 +0000 (10:25 -0400)] 
sync changes from srtp upstream

11 years agoAdded additional checks that HDLC frames arriving at the T.38 gateway engine
Steve Underwood [Tue, 8 Jul 2014 12:48:15 +0000 (20:48 +0800)] 
Added additional checks that HDLC frames arriving at the T.38 gateway engine
begin with the correct 2 bytes.

11 years agoAdd new option to request TEP to be used for an outbound fax call
Patrice Fournier [Mon, 30 Jun 2014 20:34:18 +0000 (16:34 -0400)] 
Add new option to request TEP to be used for an outbound fax call

11 years agoIt was impossible to disable ECM with T.38 in gateway mode
Patrice Fournier [Mon, 30 Jun 2014 20:30:26 +0000 (16:30 -0400)] 
It was impossible to disable ECM with T.38 in gateway mode

11 years agoDisabling Require timer for T.38 re-Invites cause problems
Patrice Fournier [Mon, 30 Jun 2014 20:24:51 +0000 (16:24 -0400)] 
Disabling Require timer for T.38 re-Invites cause problems

Disabling Require timer for T.38 re-Invites tells the remote side it
doesn't need to refresh the session but FreeSwitch will still terminate
the call if the remote session doesn't refresh.

11 years agoLets get loud
Brian West [Mon, 7 Jul 2014 19:51:43 +0000 (14:51 -0500)] 
Lets get loud

11 years agoFix BLF confusion when dealing with the same extension in multiple domains
François [Sun, 6 Jul 2014 14:57:50 +0000 (16:57 +0200)] 
Fix BLF confusion when dealing with the same extension in multiple domains

11 years agoupdated directory using asr/tts
Brian West [Sat, 5 Jul 2014 14:22:59 +0000 (09:22 -0500)] 
updated directory using asr/tts

11 years agoUpdate to use PocketSphinx 0.8
Brian West [Sat, 5 Jul 2014 14:14:02 +0000 (09:14 -0500)] 
Update to use PocketSphinx 0.8

11 years agoRefactor the curl PUT read callback
Travis Cross [Fri, 4 Jul 2014 07:47:04 +0000 (07:47 +0000)] 
Refactor the curl PUT read callback

11 years agoAdvance the data_stream data pointer
Travis Cross [Fri, 4 Jul 2014 07:43:59 +0000 (07:43 +0000)] 
Advance the data_stream data pointer

11 years agoAdd PUT support to mod_curl `curl` api and app
Travis Cross [Fri, 4 Jul 2014 01:48:13 +0000 (01:48 +0000)] 
Add PUT support to mod_curl `curl` api and app

With both the `curl` api command and application, you can now ask for
data to be PUT where previously it could only be sent via POST.

11 years agoFixed a clang-3.5 missing-prototype warning and added doxygen
Kathleen King [Fri, 4 Jul 2014 01:51:14 +0000 (18:51 -0700)] 
Fixed a clang-3.5 missing-prototype warning and added doxygen
documentation for switch_channel_set_presence_data_vals. #doxygen

11 years agoFixed dead code.
Kathleen King [Fri, 4 Jul 2014 00:03:01 +0000 (17:03 -0700)] 
Fixed dead code.

While reviewing code I noticed some dead code. It was not possible to
default to the channel variable because the parameter could not be
both full and empty.

If the parameter is not a non zero length string then the code looked
like it was intending to default to the channel variable
'presence_data_cols'. If neither of these are the case it is a noop.

By enabling the dead code, you now have access to set the
'presence_data_cols' in the dialplan or scripts like lua.

11 years agoupdate image
Anthony Minessale [Thu, 3 Jul 2014 22:58:55 +0000 (03:58 +0500)] 
update image

11 years agoMerge pull request #2 in FS/freeswitch from ~KATHLEEN.KING/freeswitch-fork:doxygen...
Mike Jerris [Thu, 3 Jul 2014 21:45:52 +0000 (16:45 -0500)] 
Merge pull request #2 in FS/freeswitch from ~KATHLEEN.KING/freeswitch-fork:doxygen-switch_apr.h to master

* commit '64fc3f7934888175b80e0cdd3a065d717d0a9014':
  Changed the function parameter name in the function definition to match the updated parameter name in the function declaration. #doxygen
  Changed the variable name for clarity.

11 years agoFixed trucation of value warning.
Kathleen King [Thu, 3 Jul 2014 21:41:24 +0000 (14:41 -0700)] 
Fixed trucation of value warning.

There was a parameter mismatch between abs(), which expects an int,
and atol() which returns a long. Since max_drift is defined as an int,
there is no need to pars q as a long rather than an int.

11 years agoChanged the function parameter name in the function definition to
Kathleen King [Thu, 3 Jul 2014 21:19:33 +0000 (14:19 -0700)] 
Changed the function parameter name in the function definition to
match the updated parameter name in the function declaration. #doxygen

11 years agoRemoved a useless called to abs.
Kathleen King [Thu, 3 Jul 2014 20:17:12 +0000 (13:17 -0700)] 
Removed a useless called to abs.

Clang 3.5 reported the following error: error: taking the absolute
value of unsigned type 'unsigned int' has no effect
[-Werror,-Wabsolute-value]

Subtracting unsigned variables will never be negative and will either
be the small expected value or will wrap to a very big value. This
code is trying to determine if the difference between these timestamps
is greater than 16000.

The variables last_write_ts and this_ts deal with timestamps. In the
normal case this_ts will be a larger timestamp than
last_write_ts. This change will maintain the intended behavior of
reseting the video if the difference is larger than
16000 and in the abnormal case this value would wrap and still exceed
the 16000.

11 years agoRemoved an autological-pointer-compare from src/switch_utils.c.
Kathleen King [Thu, 3 Jul 2014 19:12:22 +0000 (12:12 -0700)] 
Removed an autological-pointer-compare from src/switch_utils.c.

Building with Clang 3.5 gave the following warning: error: comparison
of array 'iface_out.sin6_addr.__in6_u.__u6_addr8' equal to a null
pointer is always false [-Werror,-Wtautological-pointer-compare]

This is a problem because as it is written the check will never be
true. A pointer to a structure within a structure will never be null. The
intention was either to null check the pointer or to check if the IP
address itself was not zero.

From context in the code this appeared to be a pointer null check so I
removed it.

11 years agopreserve timer name for recovery
Anthony Minessale [Thu, 3 Jul 2014 18:50:14 +0000 (23:50 +0500)] 
preserve timer name for recovery

11 years agoFixed trucation of value warning.
Kathleen King [Thu, 3 Jul 2014 17:53:19 +0000 (10:53 -0700)] 
Fixed trucation of value warning.

There was a parameter mismatch between abs(), which expects an int,
and atol() which returns a long. Since max_drift is defined as an int,
there is no need to pars q as a long rather than an int.

11 years agoupdate ico
Anthony Minessale [Thu, 3 Jul 2014 17:36:08 +0000 (22:36 +0500)] 
update ico

11 years agoclean up demo a little
Anthony Minessale [Thu, 3 Jul 2014 17:20:30 +0000 (22:20 +0500)] 
clean up demo a little

11 years agoclean up demo a little
Anthony Minessale [Thu, 3 Jul 2014 17:14:17 +0000 (22:14 +0500)] 
clean up demo a little

11 years agoMerge pull request #4 in FS/freeswitch from ~KATHLEEN.KING/freeswitch-fork:doxygen...
Mike Jerris [Thu, 3 Jul 2014 17:03:04 +0000 (12:03 -0500)] 
Merge pull request #4 in FS/freeswitch from ~KATHLEEN.KING/freeswitch-fork:doxygen-switch_regex.h-2 to master

* commit '35755d37d5dd47c2f8470ca9a294b55f124ee1e7':
  Fixed parameter name of variable in switch_regex_match_partial. #doxygen

11 years agotest commit hooks, whitespace change only
Ken Rice [Thu, 3 Jul 2014 13:50:04 +0000 (08:50 -0500)] 
test commit hooks, whitespace change only

11 years agoMerge pull request #1 in FS/freeswitch from ~ARESKI/freeswitch:code-style to master
Ken Rice [Thu, 3 Jul 2014 13:42:15 +0000 (08:42 -0500)] 
Merge pull request #1 in FS/freeswitch from ~ARESKI/freeswitch:code-style to master

* commit 'b0e86e65cf16f106a4254e10ab8f8a984b3699db':
  misc of python pep8 fixes

11 years agoreswig
Brian West [Thu, 3 Jul 2014 13:09:18 +0000 (08:09 -0500)] 
reswig

11 years agomisc of python pep8 fixes
areski [Thu, 3 Jul 2014 11:46:56 +0000 (13:46 +0200)] 
misc of python pep8 fixes

11 years agoChanged the variable name for clarity.
Kathleen King [Thu, 3 Jul 2014 03:10:31 +0000 (20:10 -0700)] 
Changed the variable name for clarity.

11 years agoFixed parameter name of variable in
Kathleen King [Thu, 3 Jul 2014 00:37:23 +0000 (17:37 -0700)] 
Fixed parameter name of variable in
switch_regex_match_partial. #doxygen

11 years agoincrease memory for buffers to hold new decoded stereo codecs
Anthony Minessale [Wed, 2 Jul 2014 21:55:21 +0000 (02:55 +0500)] 
increase memory for buffers to hold new decoded stereo codecs

11 years agofix jb endless loop of missing packets
Anthony Minessale [Wed, 2 Jul 2014 21:16:32 +0000 (02:16 +0500)] 
fix jb endless loop of missing packets

11 years agoUpdate SubmittingPatches doc for JIRA smart commits
Travis Cross [Wed, 2 Jul 2014 18:37:11 +0000 (18:37 +0000)] 
Update SubmittingPatches doc for JIRA smart commits

11 years agoonly count unmuted channels in positioning
Anthony Minessale [Wed, 2 Jul 2014 17:57:23 +0000 (22:57 +0500)] 
only count unmuted channels in positioning

11 years agovs2010 reswig
Jeff Lenk [Wed, 2 Jul 2014 17:46:26 +0000 (12:46 -0500)] 
vs2010 reswig

11 years agoexpand table width a bit
Anthony Minessale [Wed, 2 Jul 2014 17:18:44 +0000 (22:18 +0500)] 
expand table width a bit

11 years agoexpand positioning api usage
Anthony Minessale [Tue, 1 Jul 2014 23:49:16 +0000 (04:49 +0500)] 
expand positioning api usage

11 years agoFS-6637 try this instead
Anthony Minessale [Tue, 1 Jul 2014 18:59:40 +0000 (23:59 +0500)] 
FS-6637 try this instead

11 years agoFS-6637 nevermind, this is already in the code
Anthony Minessale [Tue, 1 Jul 2014 18:53:45 +0000 (23:53 +0500)] 
FS-6637 nevermind, this is already in the code

11 years agoFS-6637 try this patch
Anthony Minessale [Tue, 1 Jul 2014 17:31:55 +0000 (22:31 +0500)] 
FS-6637 try this patch

11 years agofs_cli: fix compiler error on CentOS 6 caused by recent short uuid logging change
Chris Rienzo [Mon, 30 Jun 2014 21:30:59 +0000 (17:30 -0400)] 
fs_cli: fix compiler error on CentOS 6 caused by recent short uuid logging change

11 years agoAdd support for 16-byte auth tag for AES GCM mode.
jfigus [Thu, 8 May 2014 17:34:53 +0000 (13:34 -0400)] 
Add support for 16-byte auth tag for AES GCM mode.

11 years agoSet the algorithm member on cipher_t when allocating AES crypto instance. Apply...
jfigus [Mon, 28 Apr 2014 14:06:17 +0000 (10:06 -0400)] 
Set the algorithm member on cipher_t when allocating AES crypto instance.  Apply same fix to NULL cipher.

11 years agoAvoid buffer-overflow on short RTCP/SRTCP packets
Travis Cross [Sun, 29 Jun 2014 18:42:29 +0000 (18:42 +0000)] 
Avoid buffer-overflow on short RTCP/SRTCP packets

In `srtp_unprotect_rtcp()` we are not validating that the packet
length is as long as the minimum required.  This would cause
`enc_octet_len` to underflow, which would cause us to try to decrypt
data past the end of the packet in memory -- a buffer over-read and
buffer overflow.

In `srtp_protect_rtcp()`, we were similarly not validating the packet
length.  Here we were also polluting the address of the SRTCP
encrypted flag and index (the `trailer`), causing us to write one word
to a bogus memory address before getting to the encryption where we
would also overflow.

In this commit we add checks to appropriately validate the RTCP/SRTCP
packet lengths.

`srtp_unprotect_rtcp_aead()` (but not protect) did correctly validate
the packet length; this check would now be redundant as the check in
`srtcp_unprotect_rtcp()` will also run first, so it has been removed.

11 years agoAvoid buffer over-read on null cipher AEAD
Travis Cross [Sun, 29 Jun 2014 17:32:33 +0000 (17:32 +0000)] 
Avoid buffer over-read on null cipher AEAD

In the defined AEAD modes, SRTP packets must always be encrypted and
authenticated, but SRTCP packets may be only authenticated.  It's
possible, therefore, for us to end up in `srtp_protect_aead()` without
the `sec_serv_conf` bit being set.  We should just ignore this and
encrypt the RTP packet anyway.

What we are doing instead is encrypting the packet anyway, but setting
`enc_start` to NULL first.  This causes `aad_len` to underflow which
will cause us to over-read in `cipher_set_aad()`.

If we could get past that, we would try to read and write memory
starting at 0x0 down in `cipher_encrypt()`.

This commit causes us to not check the `sec_serv_conf` bit and never
set `enc_start` to NULL in `srtp_protect_aead()`.

`srtp_unprotect_aead()` does not contain a similar error.

11 years agoPrevent buffer overflow from untrusted RTP/SRTP lengths
Travis Cross [Sun, 29 Jun 2014 01:10:29 +0000 (01:10 +0000)] 
Prevent buffer overflow from untrusted RTP/SRTP lengths

When computing the start address of the RTP data to encrypt or SRTP
data to decrypt (`enc_start`), we are using `hdr->cc` (the CSRC
count), which is untrusted data from the packet, and the length field
of an RTP header extension, which is also untrusted and unchecked data
from the packet.

This value then pollutes our calculation of how much data we'll be
encrypting or decrypting (`enc_octet_len`), possibly causing us to
underflow.

We'll then call `cipher_encrypt()` or `cipher_decrypt()` with these
two values, causing us to read from and write to arbitrary addresses
in memory.

(In the AEAD functions, we'd also pollute `aad_len`, which would cause
us to read undefined memory in `cipher_set_aad`.)

This commit adds checks to verify that the `enc_start` we calculate is
sane based on the actual packet length.

11 years agoFix misspelling in comment
Travis Cross [Sun, 29 Jun 2014 00:13:30 +0000 (00:13 +0000)] 
Fix misspelling in comment

11 years agoFS-6540 This should proxy all refers now, the param has changed to proxy-refer
Anthony Minessale [Mon, 30 Jun 2014 17:53:22 +0000 (12:53 -0500)] 
FS-6540 This should proxy all refers now, the param has changed to proxy-refer

11 years agofix volume adjustments in stereo conference
Anthony Minessale [Mon, 30 Jun 2014 18:39:42 +0000 (23:39 +0500)] 
fix volume adjustments in stereo conference

11 years agosome refactoring to make new stereo stuff more stable
Anthony Minessale [Mon, 30 Jun 2014 18:00:45 +0000 (23:00 +0500)] 
some refactoring to make new stereo stuff more stable

11 years agotweaks
Anthony Minessale [Sun, 29 Jun 2014 23:39:46 +0000 (04:39 +0500)] 
tweaks

11 years agocheck for member channel
Anthony Minessale [Sun, 29 Jun 2014 20:59:16 +0000 (01:59 +0500)] 
check for member channel

11 years agoCheck for too many SRTP errors before warning
Travis Cross [Sun, 29 Jun 2014 20:40:49 +0000 (20:40 +0000)] 
Check for too many SRTP errors before warning

We're checking whether we've hit the warning threshold before checking
whether we should just end the call.  This causes an off-by-one error
where we take one SRTP error more than intended.

This commit reverses the order of the tests.

11 years agoadd openal code to mod_conference
Anthony Minessale [Sun, 29 Jun 2014 20:23:03 +0000 (01:23 +0500)] 
add openal code to mod_conference

11 years agocheck for proper openal symbol to enable functionality
Anthony Minessale [Sun, 29 Jun 2014 20:22:46 +0000 (01:22 +0500)] 
check for proper openal symbol to enable functionality

11 years agoadd 44.1khz codec defs
Anthony Minessale [Sun, 29 Jun 2014 20:22:16 +0000 (01:22 +0500)] 
add 44.1khz codec defs

11 years agoadjust ringing var
Anthony Minessale [Sun, 29 Jun 2014 19:00:35 +0000 (00:00 +0500)] 
adjust ringing var

11 years agoFAX tweaks
Steve Underwood [Sat, 28 Jun 2014 18:11:25 +0000 (02:11 +0800)] 
FAX tweaks

11 years agoFixed updating of the modem type after a FAX ECM CTC.
Steve Underwood [Sat, 28 Jun 2014 08:43:05 +0000 (16:43 +0800)] 
Fixed updating of the modem type after a FAX ECM CTC.

11 years agoHandle SRTP_READ_ERROR in mod_rayo
Travis Cross [Sat, 28 Jun 2014 06:42:57 +0000 (06:42 +0000)] 
Handle SRTP_READ_ERROR in mod_rayo

11 years agoAllow more SRTP errors before killing call
Travis Cross [Sat, 28 Jun 2014 03:43:08 +0000 (03:43 +0000)] 
Allow more SRTP errors before killing call

In a carrier interop we saw the call get killed for SRTP failures
during a reinvite.  We're wondering if the SRTP errors may have been
transitory and if it may have recovered after a few more packets.

It's debatable whether we should kill calls at all for SRTP auth
failures; semantically the right thing to do when a MAC fails is to
ignore the packet completely.  So raising this limit to 100 packets
shouldn't do any harm.  With this change we still warn at 10 errors
and every 10 errors thereafter.

11 years agoRelay cause of hangup on SRTP failure
Travis Cross [Sat, 28 Jun 2014 01:18:50 +0000 (01:18 +0000)] 
Relay cause of hangup on SRTP failure

We hangup the channel after receiving 10 SRTP packets in a row with a
bad auth tag or that are replayed.  Prior to this commit we were
indicating a normal clearing.  When doing interop and looking first at
packet traces, this made freeswitch's behavior look surprising.  With
this commit we'll indicate more loudly what's happening.

11 years agoFix misspelled function
Travis Cross [Sat, 28 Jun 2014 00:32:41 +0000 (00:32 +0000)] 
Fix misspelled function

switch_rtp_set_invalid_handler has been misspelled as
switch_rtp_set_invald_handler going all the way back to the
beginning.  So while it's possible that someone somewhere could be
relying on this misspelling, I think it's more likely that no one has
used it much and that's why it wasn't spotted.  We don't even use it
ourselves anywhere anymore.

Introduced in commit: 828e03715f67f358671c6691c3e9c0f2a66ec6ec

11 years agoAdd editor footer to core files missing it
Travis Cross [Fri, 27 Jun 2014 23:55:27 +0000 (23:55 +0000)] 
Add editor footer to core files missing it

11 years agoAllow reincarnation from mod_sofia's shutdown-on-fail
Travis Cross [Fri, 27 Jun 2014 22:14:14 +0000 (22:14 +0000)] 
Allow reincarnation from mod_sofia's shutdown-on-fail

mod_sofia's parameter shutdown-on-fail now accepts the value
"reincarnate-now".  This will cause the switch to exit immediately
with a non-zero exit code so that the supervisor can recover the
switch.  For this to work you have to pass in -reincarnate or
-reincarnate-reexec to freeswitch.

11 years agoEnsure mod_sofia params can be unset or reset
Travis Cross [Thu, 26 Jun 2014 08:55:55 +0000 (08:55 +0000)] 
Ensure mod_sofia params can be unset or reset

This is the result of auditing each mod_sofia profile parameter to
ensure that it can be unset or reset after being set.  One use-case
for this being done correctly is so a later parameter in a
configuration file can reliably override an earlier one, which is
useful for setups with layered include files.

11 years agoAvoid leaking memory on sofia profile reload
Travis Cross [Thu, 26 Jun 2014 08:39:55 +0000 (08:39 +0000)] 
Avoid leaking memory on sofia profile reload

We were leaking the string values of apply-candidate-acl parameters.

11 years agoTerminate format_cdr json files with newline
Travis Cross [Wed, 25 Jun 2014 23:30:54 +0000 (23:30 +0000)] 
Terminate format_cdr json files with newline

In UNIX, text files by definition end with a newline.

11 years agoAllow setting format of log filename in format_cdr
Travis Cross [Wed, 25 Jun 2014 22:17:24 +0000 (22:17 +0000)] 
Allow setting format of log filename in format_cdr

This commit allows you to set a `log-file` string parameter in a
format_cdr profile.  This string is a template that may (and should!)
contain variables.  This template will be expanded and used as the
file name of the CDR to be written.  This parameter should contain
only the template for the file name itself; the path is relative to
the `log-dir`.