Distinguish between unconfirming the learned peer address and
retriggering the kernel stream. In particular we don't want to unconfirm
the sinks every time we confirmed our own peer, as that starts an
unconfirm/reconfirm loop.
Change-Id: I1f172385aefeacbc4585729bce25fbc68f04c2bd
While doing the A/B reassociation during an offer/answer exchange, we
don't (necessarily) want to remove all existing subscriptions. Instead
we cant to unsubscribe all subscribers so we don't do media forking, but
leaving existing subscriptions alone to make early media reception
possible. This mirros the old behaviour.
Change-Id: Ib9e6671ca2d23d1eb4509d7cf939015c816cc622
Multiple untagged monologues can exist at the same time which would lead
to a broken bencode dictionary. Instead use a pseudo label to
distinguish them.
Change-Id: I0f41c42df8ec17c1c4fb5cc6451ea039612e505f
We may have multiple subscribers, some of which may be dead/unused. We
don't care if we have these since we don't forward to them anyway.
possibly relevant for #1337
Change-Id: I3cded5080aa2005e9dd615cccf60bd4cba5feb7d
Set NO_KERNEL_SUPPORT when we don't actually kernelise the stream, and
use that flag when trying to pull stream stats.
probably closes#1337
Change-Id: I46af55e353d87c5afdda3c106d1f3470273105bf
The advertised address might be empty (trickle ICE) so use the FILLED
flag instead to see if the sink is eligible.
Change-Id: I114bd7400ccfcc3ecbc871bdcc5aee4e7d699816
Make sure janus_session lock is obtained first and websocket_conn lock
second, in order to prevent a possible deadlock.
Change-Id: I3db1d5cea0c0295cc10c71edd20c86ce054f520b
Warned-by: Coverity
This isn't really necessary as at this point the janus_session object is
private to the thread, but we add the locking anyway to silence the
warning.
Warned-by: Coverity
Change-Id: I0b192f2f5827ad917cf5110ce486fc6cd49e1a71
If a keyspace notification SET is received and the call already exists
as a foreign call, the call is first destroyed before being re-restored.
The call destruction involves a DEL from Redis on the "hosted DB"
number, which points to the foreign DB. This makes it impossible to then
restore the call because it's just been deleted.
closes#1308closes#1334
Change-Id: Ie895b021441b2d299f8ebb5bde1824b01e12633c
Also add a safeguard against filling the remote peer address with an
address from the wrong family
closes#1305
Change-Id: Iac18212b4d526a2f7d49a06ddcd724aa89b06060
The contents of the ->next element cannot be accessed completely lock
free as they're zeroed out during call removal. Instead grab a reference
to the linked next call before releasing the lock, and also lock the
next element before moving on. This requires a more granular locking as
not to interfere with call removal: One lock to protect the contained
call and the ->next, and another to protect the ->prev
Change-Id: I5474ea3f88e3276f93ba62a952b3be13c0c182e9
commit b0c722da69ad088a2eddced12b37c0546a514890
Author: Daniel Hauptmann <dhauptmann@sipwise.com>
Date: Mon Jul 26 15:35:51 2021 +0200
changed flag bit length
in call_interfaces.h changed bit length of reuse_codec from 0 to 1
commit 0313a747532d5987f25fa9edb202aa460bf98dd1
Author: Daniel Hauptmann <dhauptmann@sipwise.com>
Date: Mon Jul 26 15:29:20 2021 +0200
inversed reuse_codec logic
in test-transcode.c and call.c, reuse_codec = 0 (default) will now result in using codec_store_populate instead of codec_store_populate_reuse
commit b876bd686bd30df21a5962aca16fc1c85574f554
Author: Daniel Hauptmann <dhauptmann@sipwise.com>
Date: Mon Jul 26 15:18:19 2021 +0200
adding option to minimalize changes in the codec_store_population
added function codec_store_populate_reuse in codec.c which replaces codec_store_populate but makes fewer changes to the GLists with the old and new codecs
added flag to enable this feature (disabled by default)
commit 6fd0b701c9589b2fae00300801e02a9b5cc397ab
Author: Daniel Hauptmann <dhauptmann@sipwise.com>
Date: Mon Jul 26 14:44:42 2021 +0200
Added Option to minimize change in the codecs
In codec.c added function to populate codec store with the fewest changes between the old and new GList which contains the codecs.
Added new testroutine in test-transcode.c line 1500
Added flag to call_interfaces.h to optionally enable this feature
Change-Id: If58d9a07d114b05dfb75553a87eb4372ae949fbb
commit 3bf554a8fbae7e948343699f40d935693618b764
Author: Daniel Hauptmann <dhauptmann@sipwise.com>
Date: Fri Jul 23 13:58:02 2021 +0200
changing codec-exchange behaviour
in codec.c line 3288 function codec_store_populate now doesnt empty dst and copy new codec from src to it, instead codecs from src will be appended to dst and codec from dst, which are not
being contained by src are being removed
Change-Id: Id6b7ee65595f9cc5c71ef557c7bac5ee38f97cbe
This restores backwards compatibility with existing code parsing the
output, e.g. Kamailio to report MOS stats.
Change-Id: Ibafb2a7a3cf118166ffe1cca9a62a06b63252944
This fixes a race condition: Peer sends updated SDP with new address,
but an older RTP packet from the old address is received afterwards.
Thsi triggers learning of this old address is the "correct" endpoint.
Afterwards the peer stops sending RTP until a packet to the new endpoint
is received there, which never happens because the new endpoint has been
discarded in favour of the "learned" old one.
closes#817
Change-Id: I508f465a669f03e35ddcc6e770d5e7859e57569f
commit d15fd4a547
Author: Damir Nedžibović <damir.nedzibovic@enreach.com>
Date: Tue Jul 6 14:07:42 2021 +0200
Also free the character data.
commit 8869187215
Author: Damir Nedžibović <damir.nedzibovic@enreach.com>
Date: Mon Jul 5 16:11:32 2021 +0200
Make documentation and parameters consistant.
commit 4b15aea2ee
Author: Damir Nedžibović <damir.nedzibovic@enreach.com>
Date: Thu Jun 10 15:34:00 2021 +0200
Update documentation.
commit 6ec1b3035d
Author: Damir Nedžibović <damir.nedzibovic@enreach.com>
Date: Thu Jun 10 15:33:12 2021 +0200
Do not use metadata for setting the recording file; use output_destination instead.
commit f65a76e8a3
Author: Damir Nedžibović <damir.nedzibovic@enreach.com>
Date: Wed Jun 9 15:56:02 2021 +0200
Only append file extension if skip_filename_extension is not set.
commit 92e9d7c679
Author: Damir Nedžibović <damir.nedzibovic@enreach.com>
Date: Wed Jun 9 14:12:48 2021 +0200
Rename the option to better match its usage.
commit 11128bff49
Author: Damir Nedžibović <damir.nedzibovic@enreach.com>
Date: Wed Jun 9 13:19:10 2021 +0200
Implement support for seting an output file and folder per recording.
Change-Id: I1579d62467eaf06a7aa1ac11e59dbb374f150deb
Move RTCP processing down into the egress section and run the processing
function once for each output, because the output media must be known
when processing RTCP RRs during transcoding.
closes#1298
Change-Id: I1797bef336e27a7064b9f42ab8c25f0aade02e47
Perform reverse SSRC mapping even if output media is not known, and use
the appropriate media side when looking up SSRCs for received RTCP
timestamps.
closes#1298
Change-Id: Ifa5a982163bf7b0510ffc2a92ae25995d1adb888
Complete overhaul of the codec handling code:
*) obsolete flags `asymmetric codecs`, `symmetric codecs`, `reorder
codecs`
*) support proper codec offer/answer
*) split codec manipulation (strip/offer/accept/etc) into separate
functions for clarity and better code maintenance
*) fully update codec handlers in both directions after an answer
*) explicit allocation and handling of codecs and payload types in a
codec_store object
*) improve codec matchup logic during answer
*) more explicit handling of supplemental codecs (CN/DTMF)
*) remove now obsolete hacks for handling certain use cases
Change-Id: I996705ba8fe339524c2f70e6bb0fd854f9a1f4fb
This solves problems when the same SSRC is looped through the same call
multiple times in different mono/dialogues, with different parameters.
Change-Id: I1d033cb1f012574d82b5bcbfffe11eb5f983cfd8
Transcoding should not be decided based on the name of the codec alone,
but primarily on the payload type. First the PT needs to be compared,
then the codec type must be confirmed.
closes#1289
Change-Id: I1a8bffc6d521443aba14d9b4cf1ad4d1e21f1226
The JSON context is already in an object, therefore skip the surrounding
{} enclosure.
Also don't run the global MQTT timer when MQTT is disabled.
closes#1290
Change-Id: I63c622bd339545e625ee18def33f21de2533f949
The only reason the wait time in timer threads is capped at 100 ms is to
catch the daemon's shutdown. Since these threads already use a
condition variable to be woken up, we can use the same condition to wake
up the threads during a shutdown, allowing for a much longer (possibly
unlimited) sleep time in timer threads.
Change-Id: I3b5f280a4d502176a731e8b52b1b731022c8db4f
When a new timer object is scheduled to run in the future, we only need
to wake up one thread to notify about the new object as only one thread
is needed to handle it. The other threads can remain in the previous
sleep state.
Change-Id: I92ccec61a6c5a4a3cea5c5428c85f83707cf4d64
GLib 2.68 deprecates g_memdup() in favour of g_memdup2() as the former
was using a guint second argument which could lead to integer overflows
(gsize should be used instead).
Change-Id: Iaa3e3b1ecaab00d1ac887c915397b15b7204dd11
This makes sure all objects are cleaned up during shutdown even if the
respective timer hasn't run yet.
Change-Id: I197b930f1b6e407819cc5a8c4ebd92fcef21b2cd
Some codec handlers hold references to objects held within the
statistics hash table, therefore we must free the statistics after all
codec handlers have been released.
Change-Id: I3832dbeeaaf21d4a82be2dc2c842718415fd46f7
Recent ffmpeg versions have removed AVPacket from the public ABI,
therefore deprecating av_init_packet on statically allocated objects.
The replacement is av_packet_alloc()/av_packet_free().
Change-Id: I5f93fbd19e43413c3328e143cc9147fbdf12d910
Expanding the initialiser of the contained array eliminates a stray
warning from gcc (-Wmissing-field-initializers).
Change-Id: Icef260477d0ff7acc050375ce48cc6604ee325ff
We should eventually try to reduce the amount of -Wno-* options, but for
now this is a net improvement.
Change-Id: I3bd03679acbc157c0d1b3c257a542e2eec0e5ee9
These functions can return negative values to signal errors, therefore
use ssize_t as appropriate type.
Change-Id: I810b2d40e9d6df47269af54c59940d114bb5a672
Use a packed union to access the IV parts, which otherwise force unaligned
accesses.
Warned-by: gcc ASAN
Change-Id: I25661f1b1a00e9cd7d1e24bbcc10ed103ccef323
get_ssrc_ctx() returns a new reference to the SSRC object, therefore we
must release the referece when we're done with it.
Change-Id: I0db07e4cca49a37af68d072ac6d0630c025b8809
Convert ssrc_ctx_put() and ssrc_ctx_hold() to macros to aide in
debugging reference leaks. Having them as inline functions hides the
actual location of the refcount changes.
Obsolete ssrc_ctx_get() as returning values from macros is awkward and
it was only used in two locations.
Also fix a function invocation mishap: obj_get_o() should be used
instead of __obj_get() as the latter is an internal macro.
Change-Id: Icc0d63f04b3816632fd120c1c749cafabbbfa331
This makes the type in line with string(3) functions and eliminates some
compiler warnings.
Also update the related bencode data type.
Change-Id: I7ef4024f4b5a0f737b3dbe03bcd078032395bce6
The SSRC mapping strategy can change if a re-invite disengages
transcoding, therefore we always need to update the output SSRC mapping
even if the SSRC entry has already existed, to avoid stray SSRC changes.
Change-Id: Ib6f14ede1a4e615ff5eb8372cc68bf1acdd4b6c8
(cherry picked from commit dd7acd2644)
Make sure the pointers we return from our continuous memory buffer is
always 64-bit aligned as it's used not only for strings, but also for
structs/objects, and such unaligned memory access is undefined on some
archs and flagged as such by ASAN.
Change-Id: I84cf74e4e9d203fe02507aa1190ccc6554fb36e2
Avoid accessing memory via pointers that may not be aligned, which is
undefined behaviour on some archs. Use memcpy for this purpose instead.
Change-Id: Iec6c8d15fdd7ef00896e494b69412847b637b01b
Check if the uint64_t pointer is aligned before using it. If unaligned,
go byte by byte. Unaligned pointer access is undefined on some archs and
flagged as such by ASAN.
Change-Id: I3afc80a2ddbc874a62d6930971493f8d461aa452
Since dh_auto_test doesn't execute the test suites through make
directly, but instead runs `make -s -n` and then executes the output,
the integrated build tests fail since the sub-make doesn't return an
error as it should when attempting to build with the wrong .h
alternative, resulting in always the first .h alternative being used.
Fix this by using a wrapper script instead.
Also adjust some other related minor build details.
closes#1202
Change-Id: I4b6436295c6b39117bd06df53aa5afc7118ad6a1
If we receive an SDP with a DTLS fingerprint, by default we adopt the
hash function used for that fingerprint in subsequent communication with
that peer. However, if the SDP is an answer, and we previously used a
different hash function in the offer towards that peer, then a later
re-invite offer would be sent with a different fingerprint, causing an
unexpected DTLS restart. Instead, make sure we don't change fingerprints
if one was already sent.
Change-Id: I603bb86ce2d7121556c161749ed08128dd0b63b2
We simplify the function so that we do not need to meddle with pointers
and data sizes.
Change-Id: I3e653b44b28347053cc7a1053de8220c80250816
Warned-by: lgtm
While the code is not threaded, better be future-proof and use safer
APIs, which in addition has less side-effects as it does not set the
global TZ related variables.
Also check for failures.
Change-Id: I083d2b5ad6901ac1a91d42d1ab7fe3e0989b02a0
Warned-by: lgtm
The ntohs() function returns uint16_t, which will always fit within an
int (including in excess for the signed bit) on the platform we support.
Thus the check for negative values is never going to hit.
Change-Id: Ie9d5bcfe5709c50260098f91ff111755ded72921
Warned-by: lgtm
The tracker.most array is of unsigned char, so pt which gets
unconditionally assigned will always be >= 0.
Change-Id: I8adc06fc6c65aa0e33ef6614eb77c67b0dc07517
Warned-by: lgtm
Whether a bit-field is signed or unsigned is implementation specific, so
we should be explicit about this.
Change-Id: I744df3d24bc08e95fa816ba4135f19cd3a5dcb17
Warned-by: lgtm
While these usages are not supposed to be security sensitive, it's still
best to avoid the usage altogether so that we do not need to think about
it. Add a taint macro so that this does not regress in the future.
Change-Id: Ic75861ed1b7ce9dfca4f897de8be2408204ce3cb
Warned-by: coverity
__rtp_payload_type_add_recv takes ownership of `pt` and potentially
freeds it. Move the call to the end of the code block, after we're done
using it.
Change-Id: I5591cd6b22295ddcd6801e2bbbb381bff403fea8
Avoid using an explicit test for the trickle ICE endpoint address and
instead use the TRICKLE_ICE flag.
Change-Id: I821ced7e77c798004c9d88ec072848b7f6e24a85
Make the websocket_conn_init() function return an error code, and delay
the initialization after we cannot fail. And otherwise return -1, such
as when we cannot initialize the HTTP nor SSL connection.
Change-Id: I0facd53560fdb06678d7df9775be277e5c4b2cae
Warned-by: coverity
The AUTO_CLEANUP_INIT() macro is already allocating a string for this
variable, which we were then overwriting and thus leaking.
Change-Id: If655c867c5666c6cff6311f01e617647375f6279
Warned-by: coverity
The __attempt_send() function will free the gs string on success,
so we should not free it again.
Change-Id: I74c9e41854db56bf29ae541eaeebfb97ce8eeb74
Warned-by: coverity
Make sure G.722 rate multiplies is applied in the right places
Don't trust encoders to return proper timestamps, but instead track them
explicitly based on frame duration
Change-Id: Ia9fa96cf662da97159fa170c3a3f37516889e1bd
When set to `false`, no changes at all. (default)
When set to `true`, bind only one desired family local address.
Also add info in rtpengine.pod file.
Also add log for sfd with no call.
close PR #1133
Squashed commit of:
commit b5009eff0b
Author: SPChan <shihping.chan@gmail.com>
Date: Fri Dec 11 11:40:17 2020 +0800
crypto.c: handle 12 bytes master salt length of AEAD
AEAD RFC7714 specifies 12 bytes master salt length
(section 12) but the KDF expects 14 bytes.
The erratum to RFC7714 specifies to pad on the right
by 16 bits to be compliant with the KDF.
https://www.rfc-editor.org/errata_search.php?rfc=7714
commit 30ba71f10c
Author: SPChan <shihping.chan@gmail.com>
Date: Fri Dec 11 09:31:00 2020 +0800
crypto.c: AEAD make clear which operations are for AAD
commit 3e4d87681d
Author: SPChan <shihping.chan@gmail.com>
Date: Thu Dec 10 19:38:51 2020 +0800
Implement RFC7714 ciphers: AEAD AES
Initial userspace only implementation
Change-Id: I8624b255e4fd83a5ca418655ef117c48e56e2379
If a supplemental codec was removed by the codec tracker, we must reset
any codec handlers on the opposite side that were using this output
codec.
Change-Id: I7fb9d0ec7a061682fe3b28d9ead14ac263aff7c3
Sequence of events:
1) HTTP request is being handled in worker thread by calling the handler
func() from within websocket_process().
2) Handler func generates output, queues it up, and requests a
`writeable` callback from within websocket_write_raw().
3) Main LWS thread triggers writeable callback and calls
websocket_dequeue().
4) Output is given to LWS still within the main LWS thread, and finally
lws_http_transaction_completed() is called to release the connection
and ready it for the next HTTP connection.
5) LWS internally cleans up the connection and frees the user context
(our `wc` struct).
6) The worker thread wakes up and continues to use the now invalid `wc`
in order to clean up after it has done its job. Boom.
The solution is to handle the `drop protocol` callback, which is
triggered by LWS in the main LWS thread in step 4 from within
lws_http_transaction_completed(). We call our own connection cleanup
function websocket_conn_cleanup() which blocks until all jobs are
removed from `wc` (step 6) and only then continue, allowing LWS to
safely free the struct.
Change-Id: I596a98e9b552a96aef259f4523f16fa63c287ef4
Closes#1127
Squashed commit of:
commit 23dfb952e7
Author: jakubkarolczyk <51345687+jakubkarolczyk@users.noreply.github.com>
Date: Fri Dec 4 16:17:00 2020 +0000
Update jitter_buffer.c
Changed the way of excluding RTCP from being processed by JB
commit 79f478a30d
Author: Jakub Karolczyk <jakub.karolczyk@gamma.co.uk>
Date: Wed Dec 23 13:12:44 2020 +0000
For RTCPMUX streams we need to filter out RTCP packets selectively from the stream. Original implementation just stops processing packets in JitterBuffer when the first RTCP packet is received in RTCPMUX stream
Change-Id: Iad30001a6554630a63474dc069d81aefa93a0bbb
Richard Fuchs
Stefan Mititelu
Daniel Hauptmann
Damir Nedžibović
Tinotenda Chingosho
Guillem Jover
Clarence
Julien Chavanton
Julien Chavanton
Dup4
Dennis Yurasov
Huseyin Dikme
balajeesv
SPChan
dvillaume
Jakub Karolczyk