Change the log lelvel and message to avoid misleading on devices without DECT support

- Relace the calls to ubus_free() with ast_ubus_free_context()
- Remove the definition of ubus_free_context() which has been moved to Asterisk

038e34d Fix the memory leak in blob_buf
679e002 Initialize the pointer to NULL for safety
a2cb58d Fix a memory corruption issue related to ubus_free
f5cb2e8 Add emergency attribute to the output of call_status ubus object

Fix memory leak in chan_voicemngr_request()

Merge commit 9746ae8f as specified
in feeds.conf.default of OpenWrt tag v23.05.2.

Do this merge "in the other direction", i.e. merge devel into the branch
instead of merging the branch into devel. This helps git blame attribute
code lines to the actual OpenWrt commits instead of the accidentally
squashed commit 975b3a25 (Upgrade to OpenWrt-23.05.0-rc3,
2023-10-03).

This reverts the merge-commit which was accidentally squashed and by
merging re-units the history before the bad commit with the history
after the bad commit. For details, see 4edc2c18 (Merge commit
f7fbaea4 (OpenWrt v23.05.0-rc3) into
merge-openwrt-23.05.0-rc3-properly, 2023-11-22).

This merge commit does not introduce any changes to the code. It's sole
purpose is to re-unite the histories.

Merge commit f7fbaea4 as specified
in feeds.conf.default of OpenWrt tag v23.05.0-rc3.

This merge has been done in commit 975b3a25 (Upgrade to
OpenWrt-23.05.0-rc3, 2023-10-03), but it had several issues:

1. Not a proper merge commit, thus losing all history and making future
   merges hard.
2. Things from 22.03 had been left inside the branch when merging which
   created errors when running make and feeds update due to errors in
   Makefiles. In this merge we take the 23.05.0-rc3 side and discard
   changes from 22.03.

Do this merge commit right before the bad commit had been done. After this merge
commit, the bad commit and a revert of the bad commit will be merged
into this branch to re-unite the histories. Doing it this way will make
sure that git blame will attribute lines to the proper OpenWrt commit
instead of the squashed bad commit. Later we can then merge this branch
with the history of latest devel.

This reverts commit 975b3a25.
This should have been a merge-commit, but it has been accidentally
squashed.

 * Fix a crash due to wrong handling of 200 OK to REGISTER    
 * merges the proper history of Asterisk 20.3.0 by reverting
an accidentally squashed merge and doing a proper merge commit after

 * fd273d62 : lower the log level for loading moh to debug   
 * a659b358 : cherry-picked the fix for unload mwi module during shutdown from asterisk project  

Parse RTCP SR and RR packets for inter-arrival-jitter calculations

Wenpeng Song authored 1 year ago and Yalu Zhang committed 1 year ago

Skip to use the combination of `sed` and `escape_sed_substitution` in this case. 
This fixes the incoming issue after the openwrt lift, due to the syntax and return
value for the combination usage of `sed` and `escape_sed_substitution`.

Use the same header for RTP/RTCP packets in DSP and Asterisk

Fix a bug when flash hook is pressed on the second outgoing call during ring-back state

asterisk-chan-voicemngr_2.0.22: Display caller ID for the second incoming call during call waiting if CLIP is enabled

[23.05] freeswitch: add patches for CVEs in bundled libvpx

CVE-2023-5217
CVE-2023-44488

Patches are for bundled libvpx, see [1].

[1] https://github.com/signalwire/freeswitch/issues/2258



Signed-off-by: Sebastian Kemper <sebastian_ml@gmx.net>
(cherry picked from commit 26e09b2914af221e477a65f1ec92db48b9a9979a)

Fix for busy tone playing

Wenpeng Song authored 1 year ago and Yalu Zhang committed 1 year ago

* Play a beep before hangup when the remote hangs up.
* Stop playing the congestion tone from the extension, just wait for 5s that the
  voicemngr plays the tone.
* Re-arrange a bit with the cbbs sequence, so we have 5s busy tone played from
  voicemngr when cbbs is disabled.

Fix asterisk hang during call conference start

Fix a race condition caused by temporary 401/407 response

(cherry picked from commit bfa9f3e4)

... by calling pthread_cancel instead of pthread_kill

Fix a bug that RTP/RTCP ports are not closed for non-answered incoming calls

Fix toggling between calls for 'uk' flash_spec.