The poller-per-thread feature was broken with a division by zero. Take
the opportunity to rework it and eliminate the poller_map object. Use a
simple array of pollers for media sockets, plus one global poller for
control sockets. In the regular case only one poller is created and
everything points to that poller. In the poller-per-thread case, one
poller per thread is created, plus one poller (also with its own single
thread) for control connections. All control sockets use the single
control poller, while all media sockets get assigned one poller from the
pool in a round-robin fashion.
closes#1801
Change-Id: Iae91a3e10b7206455c6df33b1a472254c700ce21
Look up the first object in the tree to determine the needed sleep time.
But instead of leaving the object in the tree for the next (post-sleep)
iteration, which requires another tree lookup, remove it from the tree
and remember it, anticipating that nothing would get added into the tree
to be scheduled sooner. This saves us from having to do another tree
lookup for each timer that runs.
Adapt the scheduling function to be able to handle this and wake up the
thread sooner if necessary.
Change-Id: I9297346bc4aa8d2f68ecb833be8f71ce62a3bbfe
If the timer thread is already scheduled to wake before the timer that
we want to schedule, then there is no need to wake it up for nothing.
Change-Id: I5e6174bb0347954dfc5d012befe31625fad9961a
This should prevent multiple threads from fighting over the same shared
structures (GTree + mutex + condition). Downside is that it may lead to
load not being perfectly balanced between threads.
Change-Id: I8127b98dcfbeafd692d74e60cdf6d60e3e572ba7
For convenience we provide extra HTTP and WS endpoints that accept a
cookie-less NG or JSON message string. Not all commands are sensitive to
retransmits and this makes it easier to query call status etc.
Change-Id: Iffbc4ef9a5fdf916a374dfdd4042c61b437d18c9
Distinguish between functions that enagage or disengage the recording
daemon and functions acting on actual recording, which eliminates the
need to deal with the flags separately.
Change-Id: Ia2d718d9e6f95d7621a2ba186c60b501f7404fe7
Store SDP session attributes in the monologue that were received instead
of the ones to be sent out. Look up the corresponding source monologue
when creating an outgoing SDP and insert that one's attributes.
Change-Id: Ic2e57db71b0e0b48b9868940d326b5a951d89e92
Make 'extmap' attributes another type of OTHER attribute, which are
printed verbatim, but instead of skipping over them when parsing them
and "strip extmap" is requested, always parse and store them and then
skip over them when printing them.
Change-Id: I9f8e9c468ea0d1b20d82ae563c334513a8c0c43e
All OTHER attributes are printed verbatim as strings, but we still want
to be able to distinguish between different types of them. Introduce an
extra enum for that purpose.
Change-Id: I24351c26a45ffbbe1dec9299faa5d0ea65cada29
Even though both true and false are set in different calls to the
function, there's no effective difference in output.
Change-Id: I0d29dbf2d8041523b6b28bb0c0270bb6b414dcb5
Make sure sdp_attributes is only used for _received_ SDP attributes, and
not ones that we want to send out. The only places where self-generated
attributes are being used is the T.38 gateway.
Change-Id: I1f035cd42a35fa250ea477700a1a88ad59efcc38
Use a settable callback function to do attribute printing. This makes it
possible to use different sources for SDP attribute output in different
situations.
Change-Id: I30be7772dd73d9ca8cb9483de53eb60734e12950
While processing SDPs, we only need to look up manipulation
instructions, which means we can just return NULL if none were given,
instead of always creating the object.
Change-Id: If187c1289eb385b23b86b3aa326634efaf0ce5ad
Create a separate smaller struct that contains only the relevant pieces
of the attribute string. Embed this into the larger `sdp_attribute`
struct, which also contains all the parsed-out information, which is
needed only during initial processing.
Functional no-op.
Change-Id: I5a70f8b7c9922c312aa03782f85e7db40f054092
Only the legacy protocols (which hardly anyone uses) don't pass a
`flags` object to the offer/answer functions. Make them also pass a
dummy `flags` object so that all the non-null checks become redundant
and can be removed.
Change-Id: Id2ec76ba307ad80958168a3a42e5679687765e22
Handle the `output-destination` flag in the same way as the `metadata`
flag, eliminating the need to do it explicitly everywhere it's relevant.
Add a few alternative spellings of `output-destination`.
Rename the member from flags to `recording_file` to make it more clear
what it is.
Add save/restore capabilities of that field to the Redis code.
Unify printing of monologue LABEL into the same function that handles
the metadata.
Update documentation to better explain this option, and use the new
clearer name.
Change-Id: I4496341013b0ccab5b1dec026cf3a1a0ea879018
Move the code generating the recording metafile prefix out of the
"externally" visible code and logically deeper into recording.c.
Move the field from the recording struct into the call struct so that it
can be directly restored from Redis, and eliminate and mostly unused
function argument in the process.
Functional no-op.
Change-Id: I31eb3cd5864dd8138fabad0bcbd08ef18571f1a9
Switch the start/stop/etc recording methods to parse out the full
message dictionary into a sdp_ng_flags like all other methods, and then
pass the sdp_ng_flags object to update_metadata_*() for a unified
recording setup function.
Functional no-op.
Change-Id: Iad1004981808c45973bfd9bc2dc6c461acc602cc
Graphite isn't able to deal with spaces in the metric names delivered to
it. Introduce special version of the command strings with spaces
replaced by underscores.
closes#1780
Change-Id: Ie8bcec5ca4f2d427e92901f6fa76b985df6e459e
In the header files follow the rules:
1. Firstly goes ifndef/define construction, then one empty row.
2. Secondly go system headers, so in angle-brackets, then one empty row.
3. If there are important pre-processor definitions, which affect
the following custom file headers, they are added next, then one empty row.
4. Thirdly custom header files, so in double quotes,
then at least one empty row.
5. If there is "xt_RTPENGINE.h", it's mentioned next, but separately,
then one empty row.
6. Then pre-processor definitions, and one empty row before the code.
In some situations it's allowed to step aside from the rules,
when inclusions are dependent on each other, so on specific sequence,
and also possibly on some inline objects definitions.
But, if possible to follow the rules, it should be done.
Change-Id: I6bec69b508653947c04e7785775373d21112eb58
Instead of using transcoding flag `_TRANSCODING`
on the monologue level, we have to use that on the media
level in order to properly reflect the level on which
transocding is being used, and also to be able to selectively
set this for specific media sessions.
Change-Id: I9a25dc7be24f80b2b6ada816448a67933c762d86
Sinks must be a list of medias, not monologues.
Refactor `call_block_silence_media()`.
Additionally: deprecate the `call_media_subscribed_to_monologue()`
because of no use.
Change-Id: Iafa4fc87d613948b84373985711e8cebfa48fad8
gcc <12 propagates the const-ness of the given value to the newly
declared variable with __auto_type, leading to silly "can't assign to
const variable" errors. Work around this with an ugly macro.
Change-Id: Ic952c094c24bd802379fc10ad19d559613b2c1d0
There's two flavours of this struct being in use, even though the
structs' signatures are the same. One contains socket_t objects, the
other contains stream_fd objects. Separate them out and be explicit
about which is which.
Change-Id: I5ef1d154cc442528149f69be2e6a02625a6c650d
Simple wrappers around glib's GHashTable functions that add type
information, instead of just using void* for everything.
Change-Id: I3e7c0f095f1f58b943ba80d1e14e763707ee1698
Allow codec_tracker_update to reference an existing codec_store. Then
when a supplemental codec type needs to be generated, make it look up
the type in the existing codec_store and re-use the existing payload
type if present instead of creating a new one. This allows payload type
numbers to remain unchanged during a re-invite.
Change-Id: I9e5edd897515a5e3eb5033aa6bbf21c8667d6133
During an offer, we update the codecs from the given list not only on
the side of the offerer, but also on the answerer's side, in order to
perform the codec answer routine during the answer phase. While doing
this, we empty out the existing list of codecs (on both sides) and
repopulate it fresh from the given list.
This can cause problems during a reverse re-invite, when the list of
codecs on the answerer's side already contained the codecs that had been
offered before. When setting up the new re-invite offer, we want to
retain codecs (and their payload types and format parameters) that were
already in place, instead of recreating a new list from scratch.
Improve this by adding a `merge_cs` option to the populating functions,
which points back to the stream_params codec_store. Codecs that would
have been removed from the codec_store during the repopulation are then
moved back into the stream_params codec_store instead. This then allows
the functions adding new codecs to the list (offer/transcode) to
reference these codecs that were previously in place, and so they can be
added back with the same options as they had existed before, instead of
recreating them from scratch.
Change-Id: I53e7ab10e9144a308a5c36be5ebfddd73c212f06
From now on the `call_subscription` concept gets
deprecated, and instead of it the `media_subscriptions`
concept gets applied.
Benefits of this change is:
- ability to subscribe one-to-multiple medias (different monologues)
- media level manipulations, without affecting whole SDP session
- no need to use medias offset, to detect proper subscription's media
- there is no need of particular medias order, they can be
subscribed to each other in any possible way
(even though RFC still requires to always have proper ordering)
Deprecated objects:
- `struct call_subscription`
- `GQueue subscriptions`
- `GQueue subscribers`
- `GHashTable * subscriptions_ht`
- `GHashTable * subscribers_ht`
Deprecated functionality:
- `__unsubscribe_one()`
- `__unsubscribe_all_offer_answer_subscribers()`
- `__unsubscribe_from_all()`
- `__subscribe_offer_answer_both_ways()`
- `__add_subscription()`
- `__unsubscribe_one_link()`
- `call_get_call_subscription()`
- `call_subscriptions_clear()`
- `call_subscriptions_free()`
Offtopic: additionally this commit adds helper func:
- `call_media_subscribed_to_monologue()`
Change-Id: Ifb44f7a1ba5b483b1472882b1b8d06444dba1727
In the `dialogue_unconfirm()` (previously `dialogue_unkernelize()`)
use media subscriptions to find subscribers/subscriptions
and correlated to them medias to unconfirm them.
This change is bound to the switch from monologue
subscriptions to media subscriptions.
Additionally: Refactor unkernelize/unconfirm in `call.c`
To provide a sense of the functions clearer,
the namings of the following function singatures
have been changed:
- `dialogue_unkernelize()` -> `dialogue_unconfirm()`
- `__monologue_unkernelize()` -> `__monologue_unconfirm()`
Change-Id: If8cc2515eb26dcc4ffed02a420514f98d43cc5c6
Add handling of SDP session level attributes for:
- `monologue_subscribe_request()`
- `monologue_subscribe_answer()`
This will be used by `janus.c` related functionality.
Change-Id: I1c50b5b9da08e7d8cb2c98eb6995d8610386d6ed
We have to move the list of attributes with following types
to the `call_media`'s `sdp_attributes` using `stream_params`:
ATTR_SSRC
ATTR_MSID
ATTR_OTHER
This is then required by the `insert_sdp_attributes()`
used by:
- `sdp_replace()` -> `replace_sdp_media_section()`
- `sdp_create()`
Additionally: move the `insert_sdp_attributes()`
to the usage of the `append_attr_to_gstring()`, which
gives SDP attrs manipulations control over attributes
being inserted into SDP. Previously not supported.
Additionally: `process_media_attributes()` takes new
argument `bool strip_attr_other`, in order to control
whether or not to strip `ATTR_OTHER/ATTR_SSRC/ATTR_MSID`.
For now there is only one user of it `replace_sdp_media_section()`,
but for later on, we can control this using this bool.
Additionally: `print_sdp_media_section()` takes new
argument `bool print_other_attrs`, in order to control
whether or not to strip `ATTR_OTHER/ATTR_SSRC/ATTR_MSID`.
This is required, because the sdp replacing/creating mechanism
needs a controllable way of adding them, users are:
- `janus_videoroom_join()`
`janus_videoroom_configure()` - always need them
- `call_publish_ng` - always needs them
- `call_subscribe_request_ng` - never needs them
(otherwise leads to a duplication of such attributes)
- `call_offer_answer_ng()` - requires them always
- `call_subscribe_request_ng()` - never needs them
(otherwise leads to a duplication of such attributes)
Change-Id: I9c83d99da1603acb55443c462797b2cd1e72477d
Stop using the call subscriptions model in the
subscribe request/answer functionality and move to
the media subscriptions.
Change-Id: I8aab2b1b4cdf9a3c5a04172c395ca509295ce0a3
For the use case with a separate base chain, we want to preempt other
firewall rules by default and insert our immediate jump rule at the
beginning, rather than at the end. Add an option to provide the other
behaviour.
Change-Id: I16171f7c028c89b94823ecc99387771be3ba9443
If we're updating the handlers for one particular source -> sink flow,
only stop/reset the handlers matching this flow.
Change-Id: I1d046f47f8d26cac47c5d0f4318498eacb6c5677
With the "foreign" flag now stored in call_flags, restoring a call from
Redis restores this flag as well, overwriting the desired flag as we had
set it through call_get_or_create().
Reverse the flag setting by taking it out of call_get_or_create() (where
it's always false anyway, except when coming from Redis) and setting it
explicitly with call_make_own_foreign() after restorting call and its
flags.
Change-Id: Ib68be2aeedfa988b7555e426fa337657e1062245
Start using the media subscriptions model
(based on newly introduced `media_subscription` objects)
in scope of:
- `redis_encode_json()`
- `json_link_medias()`
- `rbl_subs_cb()`
Change-Id: I3f7267ab156b361d7e7bec4ff91a8976a7be02ee
UDP packets sent in response to a UDP request should have the same
source address as the request's destination address.
This can be achieved with sockets bound to a specific address, but in
the case of ANY-bound sockets, we can use the PKTINFO mechanism to do
this.
Extend control_ng_process() to accept an extra socket address
corresponding to the local address to use. Extend the signature of the
callback function (to do the actual sending) accordingly.
Extend socket_sendiov() to be able to set the PKTINFO cmsg when sending
a packet.
Add socket_sendto_from() as a convenience wrapper.
Extend control_udp_incoming() to pass the address from
udp_buf->local_addr back to socket_sendiov().
Change-Id: Idd019fdcfd796098e7807427e6686d4b05de35d1
Convert each listener entity into a list.
Support a list of values for each option so that multiple
ports/addresses can be listed.
Keep previous behaviour unchanged: If ANY address is given, open
listeners for IPv4 and IPv6.
Change-Id: Ic54f28d1262f60d5e5c9d824a95e7c33ebc2aba9
Start using the media subscriptions model
(based on newly introduced `media_subscription` objects)
in scope of `hunt_ssrc()`.
Change-Id: I08c7292e0d73a486e25004358de91a3414ce886a
Introduce `media_unconfirm()` to do a selective handling
of medias to unkernelize the correlated to it streams.
This is analogue of the of the `__monologue_unconfirm()`
but for medias.
Change-Id: Id8ede0fc56412021c301e97764fd5dd070b7d484
Previous implementation assumes that we use the `call_subscription`
objects in:
- `call_offer_answer_ng()`
- `call_update_lookup_udp()`
- `call_request_lookup_tcp()`
when appealing to the `call_get_mono_dialogue()`, in order to
get the `call_subscription` objects, in order to then pass it
for usage in the `monologue_offer_answer()`, where the most important
again is to use monologue references stored inside
the given `call_subscription` objects.
Instead of using the `call_subscription`, just use `call_monologue`
objects as a base data objects for this work,
which will allow us in the coming commits to deprecate
the `call_subscriptions` based model and
get to the subscription model based on medias.
Change-Id: Ia9ee5ba66522929acbceca28854ebccd3705635a
Introduction of the basic functions required for
a handling of the media subscriptions:
- `__unsubscribe_media_link()`
- `__unsubscribe_media()`
- `__unsubscribe_all_medias()`
- `__add_media_subscription()`
- `__subscribe_medias_both_ways()`
- `__subscribe_matched_medias()`
- `call_get_media_subscription()`
- `call_totag_subscribed_to_monologue()`
- `call_viabranch_intact_monologue()`
Change-Id: Iad2af5323b2ea8a10a83064d9ee72106c1d8f9c0
Initial introduction of objects we need to handle for media
subscriptions implementation:
- `struct media_subscription`
- `media_subscribers` and `media_subscribers_ht` (quick access)
- `media_subscriptions` and `media_subscriptions_ht` (quick access)
Change-Id: Ia2c4555cd7f4bcfe149bdb9f25081e0b3078042e
All functions that create listener objects take a read-only endpoint, so
make them as const.
Remove the extra TOS argument to make all signatures the same.
Change-Id: I722c7665b192476d90dbf0ece200d0bfd34cb9eb
We have long switched to having a single primary global poller, so
there's no point in passing the poller pointer around and storing it in
associated objects. Remove all remnants of non-global pollers.
Change-Id: I5a3fd217d5de51e839e2b04fec7e23643ee83631
This is a dependency of the poller and so is needed in lib/ as well to
make it usable.
Also consolidate the type.
Change-Id: I70ec8a200d6cd65710ac93636a9495cf24c35ef4
All timers have been moved to their dedicated timer threads, making this
mechanism obsolete.
The only victim is the timeout handling for TCP control streams. Since
other TCP streams aren't using timeout handling either, and the TCP
control socket is barely used by anyone, we can live with not having a
dedicated timeout for these streams for now.
Change-Id: I83d9b9a844f4f494ad37b44f5d1312f272beff3f
The original sdp_manipulate_check's only purpose left is to check for an
attribute removal. Refactor accordingly.
Change-Id: I3be28e0a2c04461a60766afe7096ad3fb8bbe998
In order to perform a substitution we must do a lookup to the hash table
anyway. We can eliminate an extra lookup for the same key to check
whether a substitution exists by combining the two operations into one.
Change-Id: Ife8f1fbfd77ade1986a303a3a5503f4142f7de01
Operating on an empty queue is essentially a no-op, so there's no point
in doing an extra check first to see if the queue is empty. Remove this
extra check for the CMD_ADD case.
Change-Id: I3c7524cd407fb350a46bdf8d3a91014158c6bf44
Use an indexed array to generically support all possible media types
instead of hard-coding a few possible types. This makes lookup faster
and is also future-proof.
Add helper functions to allocate and lookup the appropriate object.
Change-Id: I28406b2c5a6829330f41601fa52e738d9a8ee315
Transcoding is flow-specific, so it doesn't make sense to have a flag
for it in a call_media section. Instead we use the transcoding flag set
on the call_subscription objects (on subscribers), and set/unset a flag
on the monologue struct, depending on whether any media flows (going to
subscribers) have transcoding enabled.
Change-Id: Id671d56e56a22eaa8e56f6d449770b0c7b086cea
Pass the subscription object to codec_handlers_update to eliminate the
need for a return type and the subsequent if/else.
Change-Id: I311b3e8ca14ee5090cf329163975354385cee800
We have the subscription objects available now, so we can directly set
the transcoding flag without having to do a hash table lookup.
Change-Id: I2b85f34ca4d03dfaf81d92ea252902d1ee194efd
Instead of just return the monologues involved in an offer/answer,
return the respective subscription objects that have the monologues
pointing to each other. This makes it possible to set flags within these
objects.
Change-Id: Id88e56e1bf164a75e2172d0df04316cde5d8a955
Make it possible for a looper thread function to break out of the loop
by returning an appropriate status code.
Change-Id: I22e7789270eed4bf3340e7dae941929de58700ea
Make it possible to set output-destination immediately when starting the
reording, not only afterwards.
Closes#1667
Change-Id: Idb7de770ad5a41d6ae6055784213e7983df9bfec
Move the `ice_slow_timer()` functionality to a separate thread,
so that we do the work more efficiently, and not be dependent
on the call_timer runs by poller.
Furthermore it makes more sense to keep in the `ice.c`,
since it obviously has to do with ice timing.
Additionally:
Update the test-stats due to these changes in the `call_timer()`
We have to call the `ice_slow_timer()` now explicitely from
the test-stats.c, because the `call_timer()` is not anymore
responsible for providing stats counters rate calculations.
Change-Id: I03377dd59ea71c27497e1f4d30164075f05165cd
Similarly as for the `stats_rate_min_max()`,
move the `stats_counters_calc_rate()` functionality
to the same separate thread, so that we do the work
more efficiently and not be dependent on the call_timer
runs by poller.
Furthermore it makes more sense to keep in the `statistics.c`,
since it obviously has to do with statistics.
Additionally:
Update the test-stats due to these changes in the `call_timer()`
We have to call the `stats_counters_calc_rate()` now explicitely from
the test-stats.c, because the `call_timer()` is not anymore
responsible for providing stats counters rate calculations.
Change-Id: I1682eb76e3057f0f431c27b9633717d965313a1a
To do the work more efficiently,
and not be dependent on the call_timer runs by poller,
we should move the ports iterations (stats update from the kernel)
functionality to a separate thread, to make it faster and
not be dependent on what happens in the `call_timer` at all.
Since it has nothing to do with the call timers.
As an additional benefit: we unload the `call_timer` runner.
Change-Id: I511529ce504ef3d29f4e9d6d731ffd470d78d27a
Due to multiple threads polling all sockets for read events, it's
possible for one socket to receive a read event in one thread, then
immediately receive another read event in another thread, resulting in
two threads reading packets from the same socket at the same time.
While this is perfectly valid and correctly handled by mutex etc, it can
result in packets being processed out of order. In media passthrough
scenarios which don't do sequencing this can result in packets being
reordered.
Using a simple atomic counter we can ensure that only one thread is
reading from any one socket at a time.
Relevant to #1638
Change-Id: I406491d6ae5e13e618e153ba5463fd9169636016
To do the work more efficiently,
and not be dependent on the call_timer runs by poller,
we should move the `stats_rate_min_max()` to a separate thread,
to make it faster and not be dependent on what happens in the `call_timer` at all.
Since it has nothing to do with the call timers.
Change-Id: I9a39e1b63cb8741377f5af5b2d52d4f8b428a0ad
To do the work more efficiently and not be dependent on
the `call_timer` runs by poller, we should move
the releasing of sockets to a separate thread, to make it
faster and not be dependent on what happens in the `call_timer`
at all. Since it has nothing to do with the call timers.
Since now we have two queues:
- thread scope (local): ports_to_release
- global one: ports_to_release_glob
`sockets_releaser()` uses the ports_to_release_glob,
meanwhile appending in the `call_timer()` happens using the
ports_to_release.
Change-Id: Iadd966ac895b2dd64f81269d4fdf5d83747fe0b7
No need to iterate the list to achieve what the function does. We can
simply adjust pointers and count.
Change-Id: I849817e53d859ac73c14131b54a6d05e4efeb37e
For efficiency reasons `substitute` kind of SDP manipulations
commands have been moved to the `GHashTable` instead of the `GQueue`.
Change-Id: Iec6c44109ae912ba0de440bbed6ecaee6a238b97
Add a flag to inhibit kernel mode while DTMF injection is active. Tie in
DTMF playback trigger function: Remove stream from kernel, set inhibit
flag, play DTMF, then unset inhibit flag when all events have been
injected.
Change-Id: I533ec95a676bc7edf4dd973217f8d9499f1e22b6
Support using the SSRC TS derived from received RTP packets as "encoder"
TS (the "next" expected TS) for passthrough RTP in addition to the FIFO
TS of an actual encoder.
Change-Id: I7c49c27651eb89c5349bbf290b1c0ad160f77e3b
Instead of doing two calls to the kernel every time a forwarding stream
is deleted (one to update the stats and one to delete), combine these
two into a single call using REMG_DEL_TARGET_STATS.
Change-Id: I9fcb148930c24bc866f842a50edd613fb6296de0
When the `discard-recording` flag is given in one of the commands to
rtpengine (e.g. in the `delete` command), the metafile is renamed to a
.DISCARD suffix and then deleted.
The recording daemon then, seeing the .DISCARD suffix, proceeds to
immediately close all recordings, delete the files if any, and delete
the entries from the DB.
Change-Id: I3f0cac129f2d56cbccd770d43bf434dea6c0a0db
Rename the original call_delete_branch() to call_delete_branch_by_id()
and provide a new call_delete_branch() which takes a reference to a call
object instead of a call ID.
Change-Id: Ie549e7f550fab5edc6e81472e78565e1925fae67
These are two different file paths and each belongs to its respective
recording method. Move them into the appropriate struct. Also use
g_clear_pointer() for memory bookkeeping.
Change-Id: I25a897636d4f6e8fea3c533f9ca39c90f81390f8
Defer handling of trickle ICE updates to the appropriate function, which
also handles queuing up the updates in case they're processed before the
call or call party has been fully created.
Change-Id: I2489cb0ee96fba35003765bbdd692f02caed7055
Move trickle ICE handling out of __media_init_from_flags and into a
higher-level function. This obsoletes the special magic return value to
indicate a trickle ICE failure.
All methods accepting trickle ICE updates must now explicitly call the
trickle ICE update function.
The requirement to have a full dialogue for trickle ICE updates is
removed as trickle ICE only affects one side.
Change-Id: I0850e1858876ca7bcdd39b7144b53b5a4afed53e
Moving code for handling (queuing, dequeuing) trickle ICE fragments into
ice.c, where it makes no sense. No functional changes.
Change-Id: Ib68f82e8d58efe066fdc48cd32ca9869cdeab846
Using a pointer array instead of a linked lists allows us to directly
reference a media section by index number, without having to spool into
the linked list.
No functional changes.
Change-Id: I8b0e93f0c2e9addbcb4c938894118ed4a6aec768
We have to stop using objects of `struct port_pool` (media_socket.h),
becasue a newer approach introduced for ports allocations deprecates
usage of them.
Deprecated objects:
`port_pool.last_used`
`port_pool.ports_used`
`port_pool.free_list`
`port_pool.free_list_used`
Change-Id: I70e166753da7a43cb3b6b188c83d978b7dbce046
Introduce a reworked port allocation in RTPEngine.
The goal of this rework is to:
- simplify the logic of handling free/engaged ports
- eliminate a bottle neck begotten by overcomplicated logic
- potentially resolve the issue with "ran out of ports"
under heavy loading, when still there must be ports left
in the ports pool
Change-Id: Ifd2b1565611dd3b86c474a1ea5507fc6152fc212
Avoid calling lws_write() from threads other than the service thread, as
this might not be thread-safe. Instead store the values used for the
HTTP response headers in the websocket_output, then trigger a "writable"
callback, and finally do all the lws_write() calls from the service
thread.
Reported in #1624
Change-Id: Ifcb050193044e5543f750a12fb44f5e16d4c0a08
It substitutes a specified `a=` line taken from the concerned
media attributes list. If such line has been not found,
the attributes list remains untouched.
It subsitutes one attribute at a time, so one attribute into
another attribute.
Change-Id: Ie0a48ba46a1b196fbe33b09dedc40e4498640e34
dump_packet() had the "recording" argument removed at some point, making
the macro arguments invalid and checking for non-NULL pointless.
Change-Id: I9c021f3075ca9923217eaf7193f12c95b77ae081
Track audio writes in the mix buffer to set the `active` flag to true
whenever a write occurs, which makes it possible to create the buffer in
an inactivate state and implicitly set it active on demand.
Handle the mix buffer not returning any data in the RTP sending logic
(which is what happens for an inactive buffer) by simply not sending any
packets.
Change-Id: Iaeb0f6deadb3d90020c8c62872735cc94db80504
Similar to the existing media_player, but instead of simply producing
its own standalone output media stream, the audio_player takes over the
entire media stream flowing to the receiver, including media forwarded
from the opposite side of the call, as well as media produced by the
media_player.
Change-Id: Ic34ecf08fc73b04210cfffb1a7d795b462ece5ea
Return true/false from media player run functions to signal whether
playback is finished and whether codec handlers should be reset.
Change-Id: Ieb2e0861190ad2851d986bae28becc5548e215ce
Convenience function to check whether any formats have changed, which
would indicate that a restart of the media player is needed.
Functional no-op.
Change-Id: I2bc9b57b95bb229bc4f8cfc49ca662fa724d3642
In order to improve the efficieny of the code (make lookups
faster) use GQueues for CMD_ADD manipulations and
GHashTable for CMD_REM manipulations.
This gives the following benefits:
- faster lookup (check), if any SDP command is to be applied
(no need to iterate through all values of commands, as it used to be)
- keep a sequence for CMD_ADD values given by option flag(s)
In order to keep the code more lightweight and clear,
add a separate struct ptr to queues and hash tables into
the call_interfaces header.
Change-Id: I7e45aca4062750c7b8959473edb410ed76cc04e7
New section of option flags has been introduced for SDP body
attributes manipulations.
Three levels of the SDP session are concerned:
- session level (global one)
- media level - audio
- media level - video
Three different actions are supported for now:
- add
- remove
The value of the command has a wildcard matching approach.
Other attributes apart `a=` can not be edited by this functionality.
So such headers as: `c=`, `s=`, `o=` cannot be touched.
Change-Id: I939d4582839096b2399f7ded865e91ff6eb960a4
(cherry picked from commit 3f06c18793fe95e5b070044a0291a3e1528ac6e4)
Defined in RFC 8842. If a tls-id was previously given and a the new SDP
shows a different tls-id or none at all, we must restart DTLS.
Relevant to #1585
Change-Id: I554234dfcacbd330c7a0c6aea68f24d0449cba21
There's no point in sampling Graphite-specific stats, nor in reporting
them out, if Graphite is not enabled.
Change-Id: If8014513832485f38d81b478c695391129c21dff
Distinguish between two different types of "gauge" type metrics: Actual
gauges which (at least conceptually) have a single continuous value, and
metrics which are comprised of discretely sampled values, possibly from
multiple sources.
Real gauges with continuous values don't have mean/average/deviation
values directly associated with them, as calculating these requires
sampling or some other analysis.
Sampled metrics on the other hand do have these associated values.
Clarify which function does what and where each value comes from.
Change-Id: Iff5dd844b70ff70979b1b8c84dc7734d44b3da20
Rename structs and variables to make it clear that these min/max values
are min/max per-sec rate values.
Carry mins and maxes separately from averages. This changes the meaning
of $command_ps_avg away from an "average of averages" to an actual
average, which is more accurate.
Calculate this average based on per-interval differences and interval
duration (stats_rate_min_max_avg_sample).
Side effect: As rtpe_latest_graphite_interval_start is now set in
print_graphite_data instead of in graphite_loop_run, the test now
reports a different "interval calls duration".
Change-Id: I67b1118c18ca2464a48c4836fca3cfdb4d53c898
Perform accumulation of stats only once (i.e. increasing an actual
counter) and report stats based on differences to previous values,
instead of carrying multiple stats counters for each metric and
resetting each counter to zero whenever stats are reported.
`rtpe_stats` is the global master accumulator.
`_intv` variables are intermediate and local storage for values sampled
from `rtpe_stats` at regular intervals.
`_rate` and `_diff` variables hold stats calculated from `rtpe_stats`
and the respective `_intv` variable whenever the sampling and reporting
occurs.
`stats_counters_calc_diff` is used to calculate stats as differences
between `rtpe_stats` and the last sampled `_intv`
`stats_counters_calc_rate` does the same but calculates a per-second
rate, based on a microsecond duration.
Eliminate now-useless struct global_stats_ax
Change-Id: Ic4ca630161787025219b67e49b41995204d60573
Use dedicated functions and different code paths for the different use
cases, instead of just a single function which does if/else all the
time.
Change-Id: Ic3ab928c3605cfe4a2f48889e95a3a1ddccdbbec
The SRTP decryption context is associated with the local socket. Use the
socket that a packet was actually received on for the decryption context
instead of using the one that it was expected to be received on.
Change-Id: Iddf400a440fc51b4afb370ec827f75e9626b2cfd
(cherry picked from commit 8c3452e50b7aa4f5b7122dbd7221e34143467885)
... so that the publishing session gets notified about "webrtcup" and
not the session which created the room.
Change-Id: If7b308df4afa7afb19ecaca1f743f87c3c736007
Richard Fuchs
Donat Zenichev
Dennis Yurasov
Andrii Pogrebennyk
Orgad Shaneh
Stefan Mititelu
Michael Prokop
Tom Briden