Only use the codec handler's "kernelize" flag to determine whether a
particular payload type can be handled by the kernel or not. It's
supposed to be a reliable indicator, making checking the other flags
unnecessary.
Change-Id: I5fe58eded55ea973ed555e4376ee8200f112d162
Split function into one part handling the "target" (i.e. ingress)
portion of the forwarding chain, and one function handling just the
outputs.
Change-Id: I3766da3c4bc5caee4eb6bae8978f177a83cc231a
Instead of explicitly triggering an update of subscribers for A and B
during an offer/answer, trigger it only for one side (A) and then
iterate through all subscribers and subsriptions and update them all.
This automatically triggers an update for B, as well as any other
existing subscriptions that might be affected. Use a monotonically
increading update iterator counter to track which medias have been
updated so we don't run into infinite recursion.
Change-Id: Ie2fba8ff9c5a011bbe932559ac06e1634029a091
First step to refactoring the kernelize* functions. Keep all state in a
central struct that is passed to all functions. Requires an auxiliary
bool to facilitate a case distinction in kernelize_one, which can be
removed once refactoring is complete.
Change-Id: I5de8404d55d5b320ef5cb8cd27d5b684802fa7e7
Obsolete the two flavours of packet_encoded_tx and instead distinguish
the two flavour of sequence number handling (passthrough and
self-generated) through a function pointer in the codec SSRC handler,
directly pointing to one of the two existing codec_output_rtp wrappers.
Change-Id: Ib1141413859dab37048be9b04af7a5d235b46d0a
This seems to be an acceptable and reliable way to detect RTCP
multiplexed with RTP, even if `a=rtcp-mux` wasn't advertised in the SDP.
Take the opportunity to clean up __streams_set_sinks() a bit by giving
the variables better names.
Change-Id: I0cdc5e4a544641591fc2aabca12fb11bab3453f7
Use the correct media object (the source media) to obtain media-level i=
lines from. Adapt OSRTP output function to do the same.
Change-Id: I9ba4db286cf9d822b8986a7da2cb4f4aa1c69646
... and allow non-existing HT during lookup. Also reorder cleanup
sequence. This makes managing subscriptions in codec handlers possible.
Change-Id: I9c63fde2001e9dee384872577176fdf7f3bb9699
Separate out the two distinct use cases of requesting a specific port,
and requesting (one or more) random ports.
Change-Id: Ifbe7725d8638d49f620c9f4b9445342d180c8770
Add helper functions wrapping around reserve_port which directly return
a socket_port_link object. This allows us to rework
__get_consecutive_ports and release_reserved_port to eliminate mixing
different port numbers in the same list, which makes for cleaner code.
Change-Id: Ia9808f71c854f6efd0e93ce8df4527c4aea311bd
Empty subscriptions are not supported and would be an internal bug.
Remove expectations that the contents can be NULL.
Change-Id: Ibd7b0d86e7fb147135a93316575954aca546eac2
This flag is relevant to copy the state from A to B during an
offer/answer. It should only be set on one side, not on both.
Change-Id: Ie5c6de58a4ca11b37526b719b1937ffb08ebf512
Have one function to just set flags and init a single media, and another
one to set up sink pointers
Change-Id: I3207ea9f36a0e3575c2149e077e6106fd065fe1b
Directly return the just-created subscription from the respective
function. Saves a hash table lookup.
Change-Id: I42d9c75ebae2cc92619e78badb8ac594397b614b
Turns out that g_ptr_array_new_full or g_ptr_array_new_sized only
preallocates array members, but doesn't set the length.
Change-Id: I722e29a13c415f2a5ff62830dce60205f273b445
Grab source monologue and source media only once, and pass the object
down to the various functions, instead of doing the retrieval several
times in multiple places.
Change-Id: I88d019dbf9493fa0037161390fdb9efda58f5441
We know whether we're printing for monologues or for medias, so there's
no need to have a generic function signature.
Change-Id: I356747686adb34c19ba2ba4c77c2d0a77b85a364
Centralize opts setting, leave it only for
the `media_player_new()`.
Even though it's allowed them to be empty.
Change-Id: Ia347673f891c6c698909f57bdc7126fa5caf42bf
When MoH is triggered with `sendrecv` flag (so that
the recipient, the one who is put on hold, sees
the sendrecv state instead of sendonly/inactive),
we have to correctly process the answer coming back
to the MoH originator.
The originator of MoH must see:
- recvonly to his sendonly
- inactive to his inactive
Hence OA model is kept correct for the originator's leg.
Additionally: accordingly correct MoH tests.
Change-Id: Ida5f074d302c419c1e57e4fd624a55bfddae5587
If MoH originator has inactive stream
allow him to trigger media player.
This ensures that we support two
ways of putting on hold:
- sendonly
- inactive
Change-Id: I81b33184d1f1bcb0ad7c2bb3e7ada96b2c52b1ac
Add a possibility to add opts when actually
creating new media player.
This can be useful for cases with MoH implementation.
Change-Id: Ie7f9d290cd5e46063dd4095d57a521396b58bcfb
Mark the media player as used for MoH (via .opts),
before actually calling the play media
implementation.
This allows to differentiate a default
play media case from those used by MoH.
Change-Id: Ic7a60b6653a505bc856fec767f88988f624bdea0
If the receiver of a previously passthrough T.30 stream gets switched to
a generated PCM stream from a T.38 gateway, continue the sequencing of
the previous SSRC. Technically this is not necessary as the generated
PCM stream gets a new SSRC, but at least Asterisk seems to ignore this
and expect sequencing to continue, and will ignore PCM if the sequence
jump is too large.
Change-Id: Ia4656770db11f5fa1a1e9bf5bd71a0398deb1e00
Add a simple function that acts as an encryption callback to just update
the ext_seq (index) of the egress SSRC context. The kernel module
already does this, but the daemon only did it when SRTP was involved.
This now tracks egress packet indexes in all cases.
Change-Id: I9460744de55ead4b05aceb322fd8482442ff2b41
Some components (the timer thread used by the codec handlers in
particular) may hold references to components owned by the codeclib (the
.so handles in particular). Move codeclib cleanup towards the end.
Change-Id: Ic581588c27b69c025576c14a69e999c9ca5d1597
Fixes a corner case that happens when trying to delete a call created with
ng offer and no-redis-update flag and for which no ng answer was received.
In such cases, one would receive warning messages from Redis
"ERR DB index is out of range".
What happens is that on call creation redis_hosted_db defaults to -1. This
value is changed when writing to Redis, but the writing is not done for a
ng offer with no-redis-update.
Kudos go to Pawel Kuzak.
Closes#1910
Change-Id: Iad19ec8ac69b169e6352662697b17eb23e6970fb
Make sure a codec is not only known to us, but that it can actually be
used, in places where it makes sense. This is partially redundant
because ensure_codec_def_type already takes care of this, but a codec
definition may come from a different source, so it doesn't help to
double check.
Change-Id: I91af84afc2477840f1400674b2538ad8fb7746ee
Avoid repeated calls to the memory allocated for port pool list
management by picking out the list elements without freeing them,
storing them in the stream_fd object, and then returning them to the
list when the port is released.
Change-Id: I67cd5039e62e4d2965e85d7ba7f0454f08f40494
The parser currently relies on the appropriate parsing functions
(dict_iter, list_iter) to be called to advance into and back out of any
sub-objects containd within the flags string. However, if an unknown key
is encountered, no parsing functions would be invoked, resulting in the
parser not skipping over the sub-object and subsequently a stray error
message in the log. Work around this by detecting unparsed objects and
then using a dummy list_iter to return to a good parsing position.
Change-Id: I908aff8b9be8d12644a9faea0dcf3bf3535faf32
Split up the function into one part that parses out the string, and
another part that handles the individual pieces of information. Switch
to bool types and const arguments where appropriate.
Change-Id: If3ec8a048dbe1141dfc0b1150b69eb445216aad1
Switch from specialised handling of config sections (used to load
signalling templates) to a more general approach using a callback
mechanism. This allows us to add more information to the config file
while keeping the details of the underlying GKeyFile hidden. Use a typed
hash table for type safety.
Change-Id: I71ddfda0202b47df363bcc5acf1725078774f8f1
The codec `strip` option used to remove codecs not only from the
destination (offer to) monologue, but also from the source (offered by)
monologue. Change this to only affect codecs on the destination
(offering to) side. This fixes a problem of not being able to have
different `strip` options set for different branches.
As a side effect, this breaks existing use cases of using `strip` to
ignore codecs on the source monologue side. This functionality is
restored by adding `codec-ignore` in the following commit.
Change-Id: I022748fffc29caa3a8c96f514b2021f152e43686
These are potentially computed from inside each subdir, and in addition
due to what appears to be a regression in GNU make 4.4, where it is
reevaluating variables that contain $(shell) functions, many times (in
the order of thousands, this was slowing down the build, were on the
Debian amd64 build daemons it went from 5m with GNU make 4.3 to 2h40m
with GNU make 4.4. Although the bulk of the slow down has been fixed
with previous commits, the remaining optimizations are only to avoid
this potentially happening again in the future, and to reduce useless
duplicate work.
Instead of trying to cache the values from within make itself, where
programming this there is extremely painful, and does not seem to be
able to greatly reduce the number of calls, because the build system
is going to be called multiple times for different targets. Simply
externalize the generation into several shell scripts, that we call
to generate a make fragment that then we include from the various
Makefiles.
For a Debian build with GNU make 4.3, this reduces the amount of total
pkg-config calls from around ~1600 to 128, for dpkg-buildflags from
~1100 down to 6, and for dpkg-parsechangelog from ~56 to 17, but the
slow down is not as significant there anyway.
For a Debian build with GNU make 4.4, this reduces the amount of total
pkg-config calls from around ~2600 to 128, for dpkg-buildflags from
~2800 down to 6, and for dpkg-parsechangelog from ~350 to 21.
For a Debian build with GNU make 4.4, this reduces the build time
on this system from 2m10s to ~ 1m30s.
Change-Id: I427d0ea5106dc6ed1ff9e664ccdba2fa0725b7d0
Starting with GNU make 4.4, build time have massively regressed
where before they would take 5m on amd64 now can take 2h40m. While this
seems clearly broken, the release notes are filled with notices for
breaking changes, and in particular the one for passing all make
variables down to the invoked programs executed via the «shell» GNU make
function, so it is not clear what is expected breakage and what is not.
This has been reported in Debian, but not yet upstream, and while it
seems like a clear regression, it's not clear what will be the upstream
take on it. For now apply workarounds that do not change semantics, and
which do not regress with older GNU make versions.
Use the GNU make «origin» function instead of «?=» which defaults to
defining a variable as a recursive one. Coerce already defined variables
into simple ones to avoid GNU make re-evaluating these variables for
each «shell» function invocation.
Ref: https://bugs.debian.org/1092051
Change-Id: I076fc05dd616918473a22e7e942fecfdc9851d47
Add a semicolon only for the last switch statement
going in a row.
Fixes this Coverity Scan detected defect:
*** CID 1621493: Control flow issues (MISSING_BREAK)
/daemon/call_interfaces.c: 1063 in call_ng_flags_flags()
1057 out->allow_no_codec_media = 1;
1058 break;
1059 case CSH_LOOKUP("allow-transcoding"):
1060 case CSH_LOOKUP("allow transcoding"):
1061 out->allow_transcoding = 1;
1062 break;
>>> CID 1621493: Control flow issues (MISSING_BREAK)
>>> The case for value "15" is not terminated by a "break" statement.
1063 case CSH_LOOKUP("always-transcode"):;
1064 case CSH_LOOKUP("always transcode"):;
1065 static const str str_all = STR_CONST("all");
1066 call_ng_flags_esc_str_list((str *) &str_all, 0, &out->codec_accept);
1067 break;
1068 case CSH_LOOKUP("asymmetric"):
Change-Id: I4d6bc56d8149acd2d0d69f122e420c5a8c6ae2e1
Deleting a call can have the side effect of changing the active Redis DB
on the writing instance. Restore the correct DB afterwards if needed.
Fixes#1905
Change-Id: I12dae767ffa5d25703e024d2ec59aa21ba9da101
A call that gets created but then doesn't get initialised would have its
Redis DB left at zero. At destruction it would then try to switch to DB
zero. Fix this by using an appropriate initial value.
Closes#1905
Change-Id: I852e48c5a06b732b37d2ccd5c478de4760aacd4e
If keyspace notifications are used at all, the respective objects and
threads must be created during startup. This requires at least some
keyspace to be configured. To support usage without any keyspaces
initially (and add them during runtime), add a special case (set
keyspace to -1).
Convert all keyspace variables to signed ints. Ignore negative keyspace
numbers where appropriate. Support Redis endpoint addresses without
database number.
Fixes#1902
Change-Id: I45a3c87bc515f9b14e64ec1ec0906dde271b5f8d
If keyspace notifications are not configured at all, bail with an error
when there is an attempt to add a new keyspace or remove one. We get a
segfault otherwise.
Closes#1902
Change-Id: Ie47cf5f7762792aabe38476739f0dcf9927787ce
A matching payload type number in an answer might not be enough to
establish compatible codecs. Also check the format parameters. Reject
and ignore answer codecs that are not compatible.
Change-Id: I12a1287216886926ec4b3c704029c923f815b429
Instead of lumping all strings of a file into a single hash function,
split them up into their respective sections, as there is no point in
matching against strings that aren't part of the switch statement. This
should give a bit of a performance boost.
Nested hash sections require special handling as the wrapper script
isn't smart enough to hande them automaticaly.
Change-Id: I74863dfe6ca412d58101d37f9c9c85078826f1a4
HTTP/WS init creates worker threads, which would be terminated by a fork
to background. Reverse the order.
Closes#1896
Possibly also relevant to #1895
Change-Id: I30b61e07ad3bed41b6b241e8943ed479277c1474
The media blob is owned by the call, so we need to hold a reference to
the call until decoding is finished.
Change-Id: I6ed0d35edb485aad2811b287706ef101c4aa2c3f
Create separate function which only does the reading from DB and returns
the blob. Refactor __media_player_add_db to use this new function.
This slightly changes string allocation semantics: Instead of leaving
the string in the DB buffer and then letting
__media_player_add_blob_id() do the allocation and duplication, we now
immediately duplicate the string into the call's memory arena and
immediately free the DB buffers. __media_player_add_blob_id's
duplication then turns into a no-op thanks to call_ref().
Change-Id: I4360c7b0e8ec3c9b68cfc5b34b48115d8f4f89b0
Doing a simple prefix string match is not enough as there may be
keywords that are equal to a prefix of another keyword. Make sure the
keyword is followed by a space or newline or end of line. This also
removes the burden of having to skip over leading spaces in the CLI
handlers.
Change-Id: I118a08e1b33d9bf592d3d80101e4f0634401adbd
Use the call's memory arena to store strings (in particular the media
blob), and use intrusive allocation storage for the blob str itself.
Change-Id: I3a3d480869d61a2e66a24e88ad135bacca947072
This causes a termination with extreme prejudice. Useful for debugging
memory leak issues as it bypasses shutdown cleanup.
Change-Id: Ic39c5ad4aa5a4f9dc28274732bb736956a47e5bc
Richard Fuchs
Donat Zenichev
Michal Hajek
Lucian Balaceanu
Guillem Jover
Dennis Yurasov