To support scenarios with mismatched media sections between two sides of
a call that aren't based on explicit SDP manipulations, attempt to
lookup the peer media based on an existing subscription also for offers.
Add extra checks (media type, media ID, monologue pointer) to make sure
that the obtained media object is actually the correct one.
Use a fallback on the existing index-based lookup in case no matching
media can be found.
As a side effect, the numbering and ordering of media objects created
during offer/answer is now reversed as the function is called in a
different order. Adapt tests as needed.
Change-Id: Ia5351f7d4fed03098c6be6bea3d4c74f52b8a76b
The `answer` processing empties out the list of codecs and leaves only
those that were accepted in the answer. Side effect of this is that if
another answer with a different list of codecs comes through, them the
codec-accept function is missing the original list of offered codecs and
can yield an incorrect result.
Fix this by storing a copy of the offered codecs at the end of the
`offer` processing, and then restore this list at the beginning of each
`answer` message.
Change-Id: I3c714e80689f3c5689637cc7d1eb2f203c292a15
In many transcoding scenarios resampling is not actually required. We
can shortcut the operation by just returning the original frame, instead
of a cloned and newly allocated one. We just need to distinguish between
the cases to determine whether the frame returned by resample_frame()
needs to be freed.
Change-Id: I3f36a46bd3b967f140c8353119fdb24ad8363c15
Some compilers seem to think that there is some uninitialised usage
here. Work around this.
Closes#1891
Change-Id: Ic97a4b589fd2a0c33418a209557b7ce29009c7bf
If recording-method is pcap, then proc is zero-initialized, so
meta_filepath is empty. This shows many logs such as:
[core] Failed to open recording metadata file '(null)' for writing: Bad address
Prevent them by returning earlier.
Closes#1889
Change-Id: Ifeb3038f9a5c0bd6a9b7dd92842dd61783b40263
Use an object with local storage for sdp_address instead of just setting
a pointer. This makes it possible to put a different address.
Change-Id: I92e69b6a9eba9d0caa3853b3dcb134951d76b145
For some reason gcc 10.2.1 complains about this instance with
control_ng_flags_parser.c: In function ‘parse_transports’:
control_ng_flags_parser.c:299:2: warning: missing initializer for field ‘remainder’ of ‘rtpp_pos’ [-Wmissing-field-initializers]
299 | call_ng_main_flags(&dummy_parser, &STR_CONST("transport-protocol"), &(rtpp_pos) {.cur = STR(val)}, out);
| ^~~~~~~~~~~~~~~~~~
In file included from ./log.h:7,
from ../lib/obj.h:105,
from ../include/control_ng_flags_parser.h:7,
from control_ng_flags_parser.c:1:
../include/types.h:48:6: note: ‘remainder’ declared here
48 | str remainder;
| ^~~~~~~~~
but doesn't complain about other instances of the same usage. Oh well.
Make it shut up.
Change-Id: I6989e3c63b075b6c8f258a824557b767123dc0f6
The opmode is already determined uniquely when processing the NG
command. There is no need to explicitly set it again elsewhere. Simply
carry along the originally determined opmode.
Change-Id: I16924264b0337423aff6763c70a4b2c6dd3687db
There is enough overlap between the two enums that it doesn't make sense
to carry along both. Unify.
Change-Id: I9de8fbdb8d78a2002d8c1b62cea57188c937d61d
Allow to-tag updates in answers only until a non-provisional response is
received. Add a flag to optionally suppress this feature.
Change-Id: I67110ede662f618ba5a355500d3d84050e6318cc
If we receive an answer from a to-tag that hasn't previously seen and no
corresponding monologue exists (created from a previous offer),
previously we would treat this is a separate and new call branch, create
a brand new monologue and dissociate the previous one. This may lead to
unexpected results as this new monologue has been created without the
same initialisation as was done for the original one, and so may be left
with incorrect or incomplete data (e.g. SRTP keys, codec information,
interface bindings, etc).
Improve this by treating an unexpected and unseen to-tag as an alias to
the already existing to-tag. Going forward both tags can be used
interchangeably to refer to the same monologue.
Add a flag to suppress this new behaviour, in case some situation is
made worse by it.
Change-Id: Ie8f838eebd50d29d3549031998a2eb1f610b04bb
When re-establishing connection to redis via hostname,
at a point of time upon new re-resolve hasn't been done yet,
use a hostname for logging of failing attempts to connect,
so that it doesn't confuse users with the older IP.
For the backwards compatibility, this will still log an IP
for those setups using the IP address and not a hostname.
Change-Id: I28f9a2a194b5009040ac8ed909a65fec1f199663
When re-connecting to the remote redis server
try to re-resolve if the redist hostname
was an FQDN and not IP address.
Change-Id: Ie80e1d1a1ea76811c54123201ad4fe8cb64fc748
Refactor media level manipulations, no funcitonal change.
Just add a separate static function that does
all the job to create fictitious (zeroed) media.
Change-Id: Ifa39e31cdeaa093710cf8b834e7e08881056212f
In order to be able to control, which media types
one wants to have in the call session,
add support of the "sdp-media-remove" flag.
Syntax:
"sdp-media-remove" : ["<media-type>", "<media-type>", ...]
Additionally, add according unit tests.
Change-Id: Ic52456f8124319992ea9ca8c161daefb1df46b59
During the monologues lookup in
`call_get_dialogue()`,
instead of relying on the very first available
other side subscription, when doing a lookup
of this side, try to look into all of them
to find available media with existing subs.
Change-Id: If232cd25f26a2bc56eb1f401ba5c2c8a13e3e463
While handling offer/answer exchanges, do:
- index based lookup for sender's side (both offers and answers)
- index based lookup for receiver's side, but for offer only
- subscription based lookup for receiver's side, but for answer
Additionally: fix "media playback after delete" test,
because each new offer/answer exchange towards new branch,
means to have new via-branch value.
Change-Id: I078b7f7e58fa1aafbd8e68662f5f14ecb15b53c0
Don't unsubscribe offer from its existing
subscriptions, because it may have multiple
branched receiver's monologues, and hence medias.
Change-Id: I5caf515f35688041b684fbea9bf05ae559d43e15
When calling `__subscribe_medias_both_ways()`,
add sender to receiver, and not the other way.
This is just for simplicity of code reading
(so kinda: from -> to).
Change-Id: I7df889b7744c1fc3c1bec5016adbe6924e5396c3
Don't add session level bandwidth for subscribe requests,
because it's anyway mixing different medias from all parties
in one SDP and cannot really apply session level (so global)
bandwidth taken from one of the medias, to all medias.
Change-Id: If8fb5358cdae897ca811a3bd565d8be055e2101e
Make it so that options only need to be managed in one place, and all
code operating on the list of them is streamlined
Change-Id: Ia53b15910ec8973635e61cad1b43eff8d536a577
under normal circumstances, the rtp sink enters this function first
and updates the reti->local.family var to AF_INET so that when the
function is called for an rtcp sink, it exits the function early.
However, if media is being blocked the rtcp sink is the first to
enter the function and does so with a NULL payload type. this NULL
check is therefore required to prevent a sefgault on the call to `assert`.
closes#1876
Change-Id: Id95cbbc9ae388f50dd51417c351f4a9ba09eb5f0
If a monologue is determined to get a new to-tag, don't just overwrite
the previous to-tag, but also save the old one as an alias. Leave the
old one in the hash table as well so that future lookups can use either
to-tag.
Change-Id: I5d2d5cc17c85ec4ca2d8a20c501d2cdb6d793b61
The `if` block ends with a goto, so the indent for the `else` block is
not needed. Remove it for readability.
Change-Id: I0ab88246cfebf105c2256d129890925706917618
Previously there were explicit calls to fflush() in some code paths.
Replace with making the file fully unbuffered in case users depended on
this behaviour.
Change-Id: If49cf6135c0d9fe41c8896d4137e204da780be31
Re-move misleading log line to another place.
In the `sdp_manipulate_remove()` function:
- true - means remove
- false - means don't remove
Correct logging accordingly.
Change-Id: I3e0c08552a0192f3985da990b5f0c96cc31ece94
Convert existing sdp manipulation helpers
to work with arrays directly, independently from
which array is actually given.
Moreover, don't use the `G_N_ELEMENTS()`,
since we already have the `__MT_MAX`
as the size of the array, so we can directly use
that instead of relying on the G_N_ELEMENTS macro
(for array size checks).
Change-Id: I38e71dab30a467b933c213c8157fa06651ebaf04
Add a getter function for retrieval of top
most media from given media.
Currently existing getter for the top most
monologue's media subscription, just rename
so it makes more sense and is better to
differentiate:
`call_get_top_media_subscription()` -> `call_ml_get_top_ms()`
Change-Id: I5bf13b69b4d5f67cd6b53018b5a16d759542cd41
Add support for simplified G.107 formula, with math changed to
fixed-point. Retain legacy formula as an option for backwards
compatibility.
Change-Id: I999fc7de7be1407876c201c251538cea72b04008
Always create the offer/answer subscriptions during an offer/answer
exchange, making sure that no other offer/answer subscriptions can
exist.
Change-Id: I95e584c24e7ba74f2ead36d0c23976ef1f2e461c
Make sure stats are always added to the `delete` message.
Fixes stats being missing in case of `delete` without from-tag.
Change-Id: I635cc87b91d3e488696f33380b12dc68a24e51ad
The pcre2 API was changed from
pcre2_substring_list_free(PCRE2_SPTR *)
to
pcre2_substring_list_free(PCRE2_UCHAR **)
in 10.43. The difference is a `const` qualifier. Work around this.
Closes#1869
Change-Id: Ib3dd3003352f6c3155bb47d69ecb7a1b02f4647a
These functions expect a `rtpp_pos` argument, and while `str` is
compatible through the union, it lacks the second member of the struct,
which ends up uninitialised. Use the correct type through temporary
objects instead.
Change-Id: I0c553bdbb31c351346746b6072f2d424113bac5a
Don't proceed with the trickle updates in
case the required room, hence a call
hasn't been found.
Fixes Coverity Scan defect:
*** CID 1600057: Null pointer dereferences (NULL_RETURNS)
/daemon/janus.c: 1679 in janus_trickle()
1673 // ufrag can be given in-line or separately
1674 sp->ice_ufrag = cand->ufrag;
1675 if (!sp->ice_ufrag.len && ufrag)
1676 bencode_strdup_str(&ngbuf->buffer, &sp->ice_ufrag, ufrag);
1677
1678 // finally do the update
>>> CID 1600057: Null pointer dereferences (NULL_RETURNS)
>>> Dereferencing a pointer that might be "NULL" "call" when calling "trickle_ice_update".
1679 trickle_ice_update(ngbuf, call, &flags, &streams);
1680
1681 return NULL;
1682 }
1683
1684
Change-Id: Ib2e293c2f99e914e3d02fe43d08160ec30892ae4
0f9dd5def2 introduced this check for `source_media`
which was only in use by `sdp_replace()` functionality.
`sdp_create()`'s one never really reaches it, even
in older implementation where `handle_sdp_media_attributes()`
was used.
So for now it has no use, just remove it.
Fixes:
*** CID 1600058: Control flow issues (DEADCODE)
/daemon/sdp.c: 2690 in print_sdp_media_section()
2684
2685 /* print sendrecv */
2686 if (!flags->original_sendrecv)
2687 append_attr_to_gstring(s, sdp_get_sendrecv(media), NULL, flags,
2688 media->type_id);
2689 else if (source_media)
>>> CID 1600058: Control flow issues (DEADCODE)
>>> Execution cannot reach this statement: "append_attr_to_gstring(s, s...".
2690 append_attr_to_gstring(s, sdp_get_sendrecv(source_media), NULL, flags,
2691 media->type_id);
2692
2693 ps_rtcp = print_rtcp(s, media, rtp_ps_link, flags);
2694
2695 if (proto_is_rtp(media->protocol)) {
** CID 1600057: Null pointer dereferences (NULL_RETURNS)
Change-Id: I2d2ffbb0c9ecc0fae8890367754fbbd6e79e9c21
Not needed any more. All strings reside in the call's memory arena, and
SDP rewriting isn't a thing any more.
Change-Id: I594b85fe26f44fa25312eb97fb65de238fc18920
Avoids allocating and duplicating the string. Should be safe as all
strings that use it are allocated in read-write memory.
Requires changing some struct members to non-const.
Change-Id: Idc81a84a1454898c76d8514666c385307043bb6f
Use provided macros for initrialise str objects everywhere instead of
manually setting .s and .len. This allows for flexibility to extend str
objects.
Change-Id: I5ebca0bc82b348301c2c8e94b974c7acbf249b82
Convenience wrapper for &STR()
This is useful so that STR() and STR_DUP() can be extended to avoid
duplicate expansion of its macro arguments.
Change-Id: Ieae170807c11b0cdd8b52fac9bab98dccdc2b951
Instead of keeping one global hash/list of all received ICE fragments,
indexed by string call ID and from-tag, move the hash/list of ICE
fragments into the call object, now indxed only by the from-tag.
This requires a call object to exist before an ICE fragment can be
stored. Change the order of operations to always create a call object
first, then parse the SDP, then check for ICE fragment processing. The
determining factor now is only whether the monologue exists.
Duplicate the raw SDP into the call's memory arena instead of onto the
heap.
ICE fragments now don't need to store the pre-parsed SDP any more, and
the trickle ICE handling code is guaranteed to always have the fully
parsed SDP information.
Strings for fragment indexes can now be allocated from the call's memory
arena.
Obsolete the extra thread used to delete old ICE fragments. This is now
done in-line with regular call timeout processing.
Change-Id: I7acbc4c52666c4cdf1c02324bf33cf0bfcb4f0d0
If we remove a session that is not the session we're looking for
(re-used session ID after a removal race), we must return the incorrect
session to the hash table.
Change-Id: Iab73e1994251e5be766406b4e537a954b0e81f07
Flag the media via stream params, when it's
trickle or non-trickle offer/answer, and,
unflag it always only when it's non-trickle.
Change-Id: Ia85ce5792684a4121e224aff1a8e941e061fe5a8
The OTHER type attributes are not used outside of sdp.c. Move the
definitions into the appropriate scope.
Rename fields to be consistent with other attributes.
No-op.
Change-Id: Id92f2df2f475db92ee5ae1f3474191266d6d196d
Use a BIO WRITE callback instead of BIO_read'ing from the BIO after each
operation. This is a more direct way to intercept data that needs to be
sent out.
Implement MTU-related BIO callbacks.
Deduct the assumed IP MTU overhead from the configured MTU during
startup.
Unlike the previous code, this does not necessarily send DTLS from the
same socket that received a message, nor to the same address that sent
one, and instead always uses the selected_sfd and ->endpoint. This may
or may not be a regression.
Closes#1806
Change-Id: I4d4456df3f378d00782cbfa64afdb2a038217e6c
Refactor inactive streams handling (non-accepted media) so,
that they get also labels being printed.
Change-Id: I6627aba77a5f391f797762202eb4f0977bc0fc4a
Just remove them and re-write commentaries so,
that they make sense taking into account we don't use
the chopper anymore.
Change-Id: I60254061d6d2a7c7b1b181c288144e27b8d2c6b5
Because currently it is directly handled by the
`generic_append_attr_to_gstring()` implementation, a couple
of functions (previously used by the sdp_replace)
are from now on obsolete and not needed anymore:
- `sdp_manipulations_subst_attr()`
- `sdp_manipulate_remove_attr()`
Just remove them.
Change-Id: Ia33218fa4c144812ff3265abafb62fd7f50f6955
Instead of returning a pointer to ps, just return void,
because previously used pointer (by sdp_replace) is
not used by any other function.
Change-Id: If2e211b98ab801061f40210b571804ff8184124f
Move content of the `handle_sdp_media_attributes()` function
into the `print_sdp_media_section()` function. With all required
refactoring. No functional change.
Change-Id: Ia1e5019165a4ea281f17ac8df712701eea772155
From now on inactive medias are handled in advance
by `handle_sdp_media_attributes()`, so all non-accepted medias
are alsways caught here and never deal with
`print_sdp_media_section()`.
Hence, deprecate `is_active` bool for it and refactor the
implementation of `print_sdp_media_section()`.
Change-Id: Ief7b1de204681373660c74c71c972b289216015c
Due to deprecation of the sdp_replace
approach just rework the `sdp_version_check()`
functionality as well as the `sdp_version_replace()` so,
that it complies with current state of things.
Chopper processing is not honored anymore.
Sessions parameter has been obsoleted, hence removed.
Change-Id: I4cee1e63c70e5ff5c3694d185a1bb841b5e6fa3a
Due to no usage just discontinue the following
functionality:
- `process_session_attributes()`
- `process_media_attributes()`
Change-Id: Ie1742aaa0155e27f94c72464d527c14c4cf03ee8
Discontinue `replace_sdp_media_section()` used before
to replace media sessions, also deprecate all subsequent
functionality:
- replace_network_address()
- synth_session_connection()
- replace_media_type()
- replace_media_port()
- replace_consecutive_port_count()
- replace_transport_protocol()
- replace_codec_list()
Change-Id: I0a455f6d105fa0a2ea61461ee92944ec990240ef
Move the SDP offer/answer model to the sdp_create
approach instead of using the sdp_replace one.
This assumes the SDP body including session level
attributes (s=, o=, t= etc.), as well as the media
attributes, are formatted using currently given
session context.
In other words, rtpengine collects all possible
information during all of the offer/answer exchanges
within the dialog, which in its turn affects each
monologue's context, from which each new SDP message
will be built up.
This approach replaces the older one, which instead
used to go through the currenty processed SDP and
just replace those attributes, that rtpengine
is required to affect, leaving the rest untouched.
Additionally: all of the existing offer/asnwer model
tests were fixed to comply with the currect change.
Most of it is just an offset of attributes, which are
stored a bit higher/lower within the same media or global
SDP session.
Also a good part of unit test fixes targets a move of
`c=` line from the session level to the media one (so
each media has from now on its own `c=` attribute). This
does discontinue a support for session level connection
information.
Change-Id: Iecb4739683d23c4f9341e8a34b71f8ca2070956c
For the sdp_create approach, currently given monologue
is actually the holder of `session_last_sdp_orig` for
currently processed direction of SDP preparation.
So this -> other, whereas other keeps the required last orig.
Change-Id: Id1bdb80ecc0401be5fd430cd4369b8588de4379b
Move the `sdp_manipulations_add()` func to work with
Gstring directly, instead of using the chopper object.
This will be needed later, when working with it
from behalf of sdp_create.
Change-Id: I2843aa0c97cb968adb450ca985efa6d7d9b2275e
When processing bandwidth attributes per session
level, always update them, independently if the
currently parsed value has been set => 0.
The problem is, that if we only set them,
when initialized to != -1, then every new offer
within existing dialog, will keep bandwidth set
for the monologue, even if the newer offer
doesn't set this anymore.
Change-Id: Ifef964798a8ce9b38a4fd5f26b5bb4d0678b8829
Because not specificially used in rtpengine,
just deprecate it. Not to be mixed with
the `label=xxx` usage via NG protocol.
This will be also required when switching to
the sdp create approach, because needs to be
delivered via generics to print out SDP.
No functional change. Tests had to be fixed accordingly,
because of the appeared offset in the list of attributes.
Additionally, tests containing coming SDP with a label,
but which are within inactive media, don't pass it further.
One single example of it is an answer case of "a=mid mixup" test.
This is because of `print_sdp_media_section()` specifics,
and only printing generics for active media sessions.
Change-Id: Iabb44adf17adb5d8e0c5bcad20f527b591b2222c
Add rtcp-fb attributes applied to all payload types
via generic attributes instead of using the chopper.
This will be required when switching to the sdp create
approach.
No functional change. Tests containing `*` payload type
had to be fixed accordingly, because of the appeared
offset in the list of attributes.
Change-Id: I9fc1d2faf578a3c533cd85e5e52afd4af2062149
The `med` is assigned to NULL, which however
doesn't serve any purpose. To fix this defect report,
just remove the redundant assignment.
Fixes the:
*** CID 1600020: Code maintainability issues (UNUSED_VALUE)
/daemon/call.c: 690 in __get_media()
684 med = g_hash_table_lookup(ml->media_ids, &sp->media_id);
685 if (med) {
686 if (med->type_id == sp->type_id)
687 return med;
688 ilogs(ice, LOG_WARN, "Ignoring media ID '" STR_FORMAT "' as media type doesn't match. "
689 "Was media ID changed?", STR_FMT(&sp->media_id));
>>> CID 1600020: Code maintainability issues (UNUSED_VALUE)
>>> Assigning value "NULL" to "med" here, but that stored value is overwritten before it can be used.
690 med = NULL;
691 }
692 if (flags->trickle_ice)
693 ilogs(ice, LOG_ERR, "Received trickle ICE SDP fragment with unknown media ID '"
694 STR_FORMAT "'",
695 STR_FMT(&sp->media_id));
Change-Id: I9871be9cab8928a8e5b6ad2a5e403b9ee3d44295
Add `break` statement. Fixes:
*** CID 1600019: Control flow issues (MISSING_BREAK)
/daemon/call_interfaces.c: 1162 in call_ng_flags_flags()
1156 out->strip_extmap = 1;
1157 break;
1158 case CSH_LOOKUP("symmetric-codecs"):
1159 ilog(LOG_INFO, "Ignoring obsolete flag `symmetric-codecs`");
1160 break;
1161 case CSH_LOOKUP("to-tag"):
>>> CID 1600019: Control flow issues (MISSING_BREAK)
>>> The case for value "155" is not terminated by a "break" statement.
1162 case CSH_LOOKUP("to_tag"):
1163 /* including the To tag in the delete message allows to be more selective
1164 * about monologues within a dialog to be torn down. */
1165 out->to_tag_flag = 1;
1166 case CSH_LOOKUP("trickle-ICE"):
1167 case CSH_LOOKUP("trickle-ice"):
Change-Id: I0195341ec70ea623ea7fece3782cfdc030ee4eaf
Add `call_ng_process_flags()` based parsing
as for other opmodes, like offer and answer.
This keeps the backwards compatibility with
the older "flags" parsing approach on the
module side, as well as adds the possibility
to parse rtpp-flags on the daemon side.
As an advantage, there is no need to use
specific local parsing for things like
to/from tags, call-id, delete-delay etc.
Additionally:
- this commit introduces flags-flags parsing
for the "fatal" flag.
However, as before is only taken into account
by the `call_delete_ng()` processing,
so no functional change.
- this commit introduces main-flags parsing
for the "delete-delay" flag, which is also
only taken into account by the `call_delete_ng()`
processing, so no functional change.
- this commit adds To-tag options flag prasing
into the `call_ng_flags_flags()` function,
and is used by `call_delete_ng()` specificially,
for cases when more specific identification of
monologues to be deleted is used.
Change-Id: Ia992e5375a2f86318d9ad193a7857dd589038eed
If both source and destination medias are non-RTP/non-SRTP,
then just consider it as pass through by default.
Change-Id: Id97a916ab236561e2ae776b54d89deae4da6f3ec
Add a function `sdp_out_original_media_attributes()`
which prints original attributes into the media, for
cases like MT_MESSAGE or ice force relay.
Change-Id: Ic5a3c1e4ed4a9ff3235fdf55252eba1e75c8381f
It's been noticed that trying to print a zero
conneciton address, on the media stream, which
actually offered non-zeroed IP source, but a
zero port, leads to a segmentation fault,
because it relies in this case on presented
`sockaddr_t` information, which is actually absent.
This happens only with sdp_create on offer/answer.
Test fixed is "a=mid on zero streams".
Change-Id: Ic2f7f671c9ded9b361d6467ca013698fdff28538
Look for the first usable (non-rejected, non-empty)
media and ps, thereby to determine session-level attributes,
if any. For that to work, it has to have `->selected_sfd`
of the packet stream.
Change-Id: Id4324864571bc78936aa16e235e3866e40860420
Richard Fuchs
Orgad Shaneh
S-P Chan
Lucian Balaceanu
Donat Zenichev
Peter Pulham