With 'no assign-tmsi', regard any TMSI as invalidated at the end of a
Location Updating procedure. Hence, avoid paging by TMSI.
When 'no assign-tmsi' is set, osmo-msc does not actively assign a new
TMSI at the end of the Location Updating. However, it stores any TMSI
identity that the MS sends in a Location Updating Request. So far, this
caused osmo-msc to use the TMSI that the MS had sent in subsequent
Paging, which goes unanswered by the MS.
(After the long standing evil twin problem regarding TMSI MI has been
fixed in recent Ifdabe0b65bffafbf7b8e5cc10e2d225d1ed1cecd, there is no
longer an evil twin risked by clearing out a TMSI.)
Related: SYS#6860 OS#4721
Change-Id: I583682d1a35a70b008d7bb2d89ba7c3109a60b21
When a subscriber first attaches by TMSI only, and later tells the IMSI
via ID Response, it may turn out that this IMSI already exists in the
VLR database. If this happens, the TMSI that the subscriber issued was
not known in the existing VLR entry, indicating that the subscriber has
in the meantime camped on a different core. Which means we can assume
that there cannot be any active connections, and the old subscriber can
be discarded, for the benefit of the new one.
(We could also discard the new one, but it is more complex to reparent
the ongoing FSMs for Compl L3 than to copy some dormant VLR state.)
In vlr_subscr_set_imsi(), check for an existing IMSI entry in the VLR.
If such exists, copy any pending Paging and auth tuple state to the new
subscriber, and discard the old one from the VLR.
In order to safely discard a vlr subscriber by force, add a new vlr_ops
function: subscr_inval(), to tell the MSC that a vlr_subscr is no longer
valid.
Upcoming patch I583682d1a35a70b008d7bb2d89ba7c3109a60b21 better clears
TMSI state from the VLR, making it more likely to hit the evil twin
situation this patch fixes; hence this is, sort of, preparation.
Related: SYS#6860 OS#4721
Change-Id: Ifdabe0b65bffafbf7b8e5cc10e2d225d1ed1cecd
Better match the pattern of sdp_audio_codecs_* instead of having
foreach_ in the front. Prepare for prepending osmo_ some day, because I
plan to move the SDP API to a separate library.
Change-Id: Ia96190e0bdb513886663be1c8c12be3b403b71c9
The comments indicating which two "members" are identical are
inaccurate. (One of them is a macro pointing at the other.)
Change-Id: Ifaa2f361db77cd0ed3ad39d6ca197195b9354ea1
Previously added codecs tests uses non-default PT number sent by MT and
adopted by MO. Also test the other direction, i.e. a non-standard PT
from MO is adopted by MT.
Related: OS#6258
Change-Id: I8fbabe242982441d676d09f4d0ed7557c8349f2c
In msc_vlr_test_call.c, allow to tell MO non-default payload type
numbers in the SDP, to verify that it adopts the other call leg's PT
numbers.
Actually apply the non-default payload type number (AMR=96 instead of
the default of 112 from codec_mapping.c) and see the effects in
msc_vlr_test_call.err.
The diff shows that, as intended, the change in payload type number
should result in modifying the MGW endpoint to change the earlier '112'
to the modified '96' used in this test.
Related: OS#6258
Change-Id: I25df2ed7ad792fbe66dfd0fbf08182c9cf6cfc5b
In msc_vlr_test_call.c, allow to tell MO non-default payload type
numbers in the SDP, to verify that it adopts the other call leg's PT
numbers.
This test differs only slightly from the first codecs test, so in this
patch add the test as a 1:1 copy of the first test. The next patch [2/2]
will then show only the difference the new test makes.
Related: OS#6258
Change-Id: I618e3cf1b412985589a0c63bd76b7a60202f17b9
In patch I8760feaa8598047369ef8c3ab2673013bac8ac8a, osmo-msc learns to
handle codec mismatches reported by MT. For simplicity, that patch cuts
short the msc_vlr codecs tests by validating only the first codec.
Now test the full list of codecs properly.
This also introduces testing the re-assignment that MO does to match
MT's codec limitations, and removes the "EXPECTED FAILURE" markers.
Related: OS#6258
Change-Id: Ib933554f826c1b4347dfa3f6c4f6fe086be8b133
It was true once, but not since "do CN CRCX first"
Ie433db1ba0c46d4b97538a969233c155cefac21c
Related: OS#6258
Change-Id: I94e430e4e5b5bf18dbb155258d82f599ada453e6
This is the last missing piece that allows osmo-msc to make good TFO
codecs choices.
Since the codec_filter, osmo-msc properly gathers codec options and
limitations. But the MO call leg still assigns a voice channel before
getting a response from the MT call leg, and is then stuck with that.
Add the capability to adjust the MO call leg's codec in case the MT side
needs a different codec for TFO.
This is only relevant for 2G; on 3G we always have AMR/IuUP.
For inter-MSC handover, keep the behavior unchanged: offer only the
currently assigned codec to the remote side. Codec-changing HO should be
equally trivial to implement, but that is for another day.
msc_vlr_test_call's codec tests are adjusted to test the new feature in
Ib933554f826c1b4347dfa3f6c4f6fe086be8b133. For now, avoid change in
these tests by validating the first codec in SDP lists only.
Related: OS#6258
Related: osmo-ttcn3-hacks I402ed0523a2a87b83f29c5577b2c828102005d53
Change-Id: I8760feaa8598047369ef8c3ab2673013bac8ac8a
To parse and handle SDP included in incoming MNCC, use rx_mncc_sdp()
everywhere. So now rx_mncc_sdp() is the single implementation for
parsing the SDP string and taking action for codecs if needed.
One current dup of this code has a fall-back to use legacy bearer cap --
absorb that into rx_mncc_sdp(), so that we now also do that fall-back
for all of the incoming MNCC that contains bcap.
This is a cosmetic preparation for implementing MO Re-Assignment to
match MT's codec limitations.
Change-Id: I94ae11654e1f88fbd64361b639a4c583836dc13e
The MGCP protocol features the 'C' (call-id) to identify which
connections belong to the same call. They may be used by MGW for
accounting or management procedures.
So far we sent the MNCC callref as call-id. Instead, add a separate
unique call_id number space. Assign a unique call_id to each
transaction.
Change-Id: I36c5f159fa0b54fb576ff8bd279928b895554793
Related: OS#4854
For all 3G calls, convert IuUP <-> plain AMR/RTP on the MSC's MGW hop
like this:
Before this patch:
hNodeB <--IuUP--> MGW@hnbgw <--IuUP--> MGW@msc <--IuUP--> other call leg
After this patch:
hNodeB <--IuUP--> MGW@hnbgw <--IuUP--> MGW@msc <--AMR--> other call leg
^
This allows, in principle, 2G to 3G calls without expensive transcoding,
like this:
hNodeB <--IuUP--> MGW@hnbgw <--IuUP--> MGW@msc <--AMR--> MGW@msc <--AMR--> MGW@bsc <--AMR--> 2G-BTS
^
(So far only proven to work with AMR-FR at 12k2.)
3G to 3G calls now look like this:
hNodeB <--IuUP--> MGW@hnbgw <--IuUP--> MGW@MSC <--AMR--> MGW@MSC <--IuUP--> MGW@hnbgw <--IuUP--> hNodeB
^
Implementatino: get rid of the shim that was put in place to still send
IuUP (VND.3GPP.IUFP) to the CN. So now, for all 3G voice, the IuUP gets
decapsulated to plain AMR/RTP at the MSC's MGW hop.
What is proven to work with this patch:
successful voice call between 2G and 3G with these conditions:
- a hNodeB that stubbornly accepts only 12k2 AMR;
- a 2G BTS configured to use only TCH/F and only FR3, with only 12k2 as
allowed AMR rate.
We have not yet seen a call working for TCH/H HR3 <-> 3G, because of the
lab hNodeB's limitation to 12k2.
Future work we probably need:
- properly request and negotiate AMR rates via SDP fmtp:mode-set.
- request more RFCIs in our RANAP RAB Assignment requests
(see I61e0e9e75e3239662846fd797532acdefa9f73dc).
- Convert IuUP to AMR already at the HNBGW's MGW?
Solving this is not part of this patch.
Related: SYS#5092
Change-Id: I386a6a426c318040b019ab5541689c67e94672a1
In msc_vlr_test_call, we fail to send the right MNCC struct for
MNCC_RTP_CREATE. We should pass a struct gsm_mncc_rtp. Fix that.
Change-Id: Ia0b3253f85c716e45f925da3f58f025af1f15ec9
In order to send the MSC's RTP endpoint IP address+port in the initial
SDP, move the MGCP CRCX up to an earlier point in the sequence of
establishing a voice call.
Update the voice call sequence chart to show the effects.
Though the semantic change is rather simple, the patch is rather huge --
things have to happen in a different order, and async waits have to
happen at different times.
The new codec filter helps to carry codec resolution information across
the newly arranged code paths.
Related: SYS#5066
Change-Id: Ie433db1ba0c46d4b97538a969233c155cefac21c
Upcoming patch 'do CN CRCX first' changes the ordering of MGCP. To
properly show the change in behavior in the msc_vlr_test_call, first
clarify which side is expected to do MGCP when.
Related: SYS#5066
Change-Id: I972e7426006e5b62f81ccfe4fa224ee9eed7a7ac
Transmit and receive full SDP information via MNCC, to accurately pass
codecs choices between the call legs.
In msc_vlr_test_call.c test_call_mt(), show that when receiving MNCC,
the codec information in SDP overrules the Bearer Cap codec information
-- we expect to still receive inaccurate Bearer Cap from e.g.
osmo-sip-connector, because we have chosen to add SDP to MNCC instead of
trying to fix the codecs represented in Bearer Cap.
For internal MNCC, the MT call leg now knows which codec the MO has
chosen and assigned.
For external MNCC, osmo-sip-connector receives SDP about our codecs
choices and sends it in SIP messages, and we also receive the full SDP
information from the remote SIP leg.
Update the SDP in codec_filter every time it is received, to always have
the latest SDP information from the remote leg.
CC MNCC
| ---ALERTING--> | add local side SDP to MNCC msg
| <--ALERTING--- | store remote side SDP
| <--SETUP-RESP- | store remote side SDP
| --SETUP-CNF--> | add local side SDP to MNCC msg
| -RTP-CREATE--> | use codec_filter, add local side SDP to MNCC msg
| <-RTP-CONNECT- | store remote side SDP
There still is one problem: when initiating MNCC, we do not yet know the
RTP address and port to be used for the CN side, because the CN CRCX
happens later. So far we send 0.0.0.0:0 as RTP endpoint in the SDP,
until the CN CRCX is done. A subsequent patch moves CN CRCX to an
earlier time, adding proper RTP information right from the start.
Related: SYS#5066
Change-Id: Ie0668c0e079ec69da1532b52d00621efe114fc2c
So far, patches have set up rtp_stream to allow setting multiple codecs,
and collected the codecs information into the codecs filter struct.
Now actually use the codecs filter result to choose a codec.
Setting up the call leg FSMs and codecs still looks rather confusing in
this patch, because this is an incremental step in a larger series. The
upcoming patch 'do CN CRCX first' clarifies this substantially.
The resulting codecs behavior is tested in upcoming patch
I879ec61f523ad4ffc69a0b02810591f7c0261ff9. (The test ideally should have
come before this patch, but my time to rework this branch is up.)
With the codecs filter in place, we are ready for sending and receiving
full SDP via MNCC, see upcoming Ie0668c0e079ec69da1532b52d00621efe114fc2c
and Ie433db1ba0c46d4b97538a969233c155cefac21c
Related: SYS#5066
Change-Id: I66e7c8c5e401f4f3a7d3d42b9525b2c6e99691d9
So far, we just forwarded the Bearer Capabilities received in MNCC from
the remote MO call leg, and omitted Bearer Cap if the remote call leg
did not provide any.
Instead, always include Bearer Cap, and compose it from the codecs
filter result. Hence the Bearer Cap is now an intersection of MS, BSS
and remote call leg, instead of just the remote call leg.
Related: SYS#5066
Change-Id: I9586221ef56352b7ce4b2604ae0dc04554145a78
Do not convert to enum mgcp_codecs, but directly pass the
gsm0808_speech_codec IE from the A interface to codecs handling.
For Iu:
- RAN side: use ran_infra.force_mgw_codecs_to_ran to keep the MGW
endpoint towards RAN on IUFP.
- CN side: introduce flag ran_msg.assignment_complete.codec_with_iuup,
so to decide whether to forward IUFP towards CN, we don't need to test
the RAN type, but use the flag from the ran_msg implementation.
In msc_vlr_tests, use the SDP codec string instead of enum
mgcp_codecs.
So far limit to intra-MSC related messaging, adjusting inter-MSC
handover follows in a separate patch.
Change-Id: Ia666cb697fbd140d7239089628faed93860ce671
Allow configuring MGW conns with multiple codecs. The new codecs filter
can have multiple results, and MGCP can configure multiple codecs. Get
rid of this bottleneck, that so far limits to a single codec to MGW.
On Assignment Complete, set codec_filter.assignment to the assigned
codec, and use that to set the resulting codec (possibly multiple codecs
in the future) to create the CN side MGW endpoint.
Related: SYS#5066
Change-Id: If9c67b298b30f893ec661f84c9fc622ad01b5ee5
Indicate in the ran_infra data structure whether a RAN needs specific
codecs to be set up on the RAN facing MGW endpoint.
This allows setting forced RAN codecs as first-class citizen in the
ran_infra data structure, instead of special cases in the code (for IuUP
on IuCS).
Will be used in subsequent commit
I37f65c36af2679ecba1040a11a9aa0eb9481d817, submitted separately for
easier readability.
Change-Id: I37f65c36af2679ecba1040a11a9aa0eb9481d817
I noticed by chance that the Assignment Complete message generated in
the test lacks a remote RTP address for the RAN side.
Make the test more realistic by adding a remote RTP address and port. It
doesn't have any bearing on the tests besides more accurately showing
RTP stream setup in the logs.
Change-Id: Ia428762a16dcc17f036d725a00e0b3767418289b
Collect either the SDP or the Bearer Capabilites in the incoming
MNCC in the new codecs filter.
So far just collect the info and do not change the behavior, using the
filter result will follow in a subsequent patch.
Related: SYS#5066
Change-Id: I84d9bbca3e4061da622b1b2fc0bde8868e7e3521
For MT call, initialize the codecs filter and apply the
Codec List (BSS Supported) from Compl L3.
Related: SYS#5066
Change-Id: I530409a64d11da48518a3dc60aa3a4e47c384663
This option should be used for any executables which are used only
for testing, or for generating other files and are consequently never
installed. By specifying this option, we are telling Libtool that
the executable it links will only ever be executed from where it is
built in the build tree. Libtool is usually able to considerably
speed up the link process for such executables.
Also take a chance to add the missing $(COVERAGE_LDFLAGS).
Change-Id: I664a9d5abed2777deee302f9d3afd1bbfde7a844
Omit "in state FOO", because LOG_TRANS() already logs the state.
Most MNCC "rx" logging was duplicated. Log "rx" only once.
If there is RTP information passed with the MNCC message, log it:
- if there is SDP, log the SDP information.
- if there is no SDP, log the legacy MNCC RTP fields, if any.
One motivation to do this is to get RTP information in ladder diagrams
generated by msc_log_to_ladder.py without the need to add udtrace MNCC
logging to osmo-msc; and also to get RTP info for internal MNCC, where
udtrace doesn't apply, because no unix domain socket is involved in
internal MNCC operation.
Change-Id: I4b916cb482ed441b508c6295de211a21c49cd5c1
When the HLR fails to return auth info and authentication and ciphering
are configured to be optional, fall back to no-auth.
This patch concludes a series of preparatory patches and implements the
actual functional change.
Related: OS#4830
Change-Id: I5feda196fa481dd8a46b0e4721c64b7c6600f0d1
Add third outcome of auth_fsm: the no_auth_info_event, which should be
dispatched when auth failed because the HLR has no auth info for this
subscriber, i.e. not because an actual auth challenge failed.
No functional change: Handling no_auth_info_event separately follows in
another patch (to allow fallback to no-auth). Feed the same
_E_AUTH_FAILURE as no_auth_info_event to still behave unchanged.
Related: OS#4830
Change-Id: I5103b1f2727f1729a5517ae359df813d50436ed3
Previous patch added the AUTH_FAILURE event, which means that the
AUTH_RES event now only signals success. Reflect that in the name.
No functional change.
Related: OS#4830
Change-Id: I7124a3591fcf36cee06d7488eeb94f9b85af5dc2
Explicitly send distinct parent events on auth success and failure. So
far determining success depended only on the data pointer passed on with
the event. Distinct events clarify the logging and the FSM code.
This prepares for a third FSM outcome to be added in a subsequent patch,
to separately signal when the HLR has no auth data.
No functional change.
Related: OS#4830
Change-Id: I02776dfe6785983f2ebe398f57867f5ceb288ba0
Since call_leg_fsm_releasing_onenter() calls immediatelly
osmo_fsm_inst_term(), it meant we couldn't receive any event in that
state because osmo_fsm disables event dispatching to FSMs being
terminated.
As a result, CALL_LEG_EV_MGW_ENDPOINT_GONE was never received and hence
call_leg_mgw_endpoint_gone() was never called, which means the
mgcp_client used in cl->mgw_endpoint was never put back to the pool.
By first freeing all the children (rtp_streams), we make sure
cl->mgw_endpoint ends up with no conns and sends us the GONE event
before we go ourselves into termination state.
Related: SYS#5987
Change-Id: I2126578c4e64c9f336e8a1f6ee98de970866b8dc
Large RAN installations may benefit from distributing the RTP voice
stream load over multiple media gateways.
libosmo-mgcp-client supports MGW pooling since version 1.8.0 (more than
one year ago). OsmoBSC has already been making use of it since then (see
osmo-bsc.git 8d22e6870637ed6d392a8a77aeaebc51b23a8a50); lets use this
feature in osmo-msc too.
This commit is also part of a series of patches cleaning up
libosmo-mgcp-client and slowly getting rid of the old non-mgw-pooled VTY
configuration, in order to keep only 1 way to configure
libosmo-mgcp-client through VTY.
Related: SYS#5091
Related: SYS#5987
Change-Id: I7670ba56fe989706579224a364595fdd4b4708ff
As part of preparation for libosmo-netif migration let's move common SMPP code
into separate build-time library and use it for both smpp_mirror and OsmoMSC
renaming the files if necessary.
While at it we also fix id/password legth limits in smpp_mirror and drop unused
fields from ESME struct.
Related: OS#5568
Change-Id: I61910651bc7c188dc2fb67d96189a66a47e7e8fb
Looking at 'perf top' of osmo-msc under load shows that there's a
significant amount of time spent in terms of locking (mutex,...)
which is useless as osmo-msc is a single-threaded application.
Unfortunately libdbi doesn't provide a mechanism to perform
sqlite3_config(), so we have to do it directly here, introducing an
explicit build-time dependency (and linkage) to libsqlite3.
Related: OS#5559
Change-Id: I5bbea90d28b6d73b64b9e5124ff59304b90a8a75
With comments, clarify the code paths where a CM Service use count has
not yet been placed on the conn (just send CM Service Reject) and where
the use count is placed (decrement count on CM Service Reject).
Place the CM Service use count slightly earlier:
- it is then correctly present when checking the mobile identity in
cm_serv_reuse_conn(), avoiding the crash reported in OS#5532.
- there is only one place incrementing the use count instead of two.
Related: OS#5532
Change-Id: I6c735b79b67108bcaadada3f01c7046e262f939b
When using 'check_PROGRAMS', autoconf/automake generates smarter
Makefiles, so that the test programs are not being compiled during
the normal 'make all', but only during 'make check'.
Change-Id: I13b519e61ca0d9ce038e8c989ddac012de4a6c61
RAB Assignment Complete contains no codec info, hence
assignment_complete.codec is not set and
assignment_complete.codec_present is false.
As a result a wrong value is passed to rtp_stream_set_codec.
This fixes osmo-msc sending "a=rtpmap:112 AMR/8000/1" during MDCX in the
RAT-side connection of the call leg after having properly sent
VND.3GPP.IUFP/16000 in CRCX.
Change-Id: Ic028d35893d29f7d72f22f82ef89695229c9b01b
This way the MGW knows it has to handle IuUP in that connection (answer
IuUP Initialization, etc.).
Depends: osmo-mgw.git 1de5ed6f979bd4c1380789c9a82f8e396f05c5f8
Change-Id: I7aca671e00ed27ac03f0d106b5a6b665a9bed4c1
Ciphering is optional in both GERAN and UTRAN, however for the later
it's *required* to enable integrity protection for the signalling.
Thus we must always send Security Mode Command in UTRAN, even in
case if ciphering is disabled (UEA0) in the configuration.
The actual decision whether to send CMC/SMC or not is taken in:
* vlr_access_req_fsm.c / _proc_arq_vlr_node2(), and
* vlr_lu_fsm.c / vlr_loc_upd_post_auth().
depending on the value returned by is_ciph_required(). Let's
rename this function to is_cmc_smc_required() and ensure that
it always returns true in UTRAN.
This change fixes the Iu test cases in ttcn3-msc-test.
Change-Id: I6205f13453eff7afbf25e013d72ae98a78fcd31b
Fixes: OS#5333
Neels Hofmeyr
Vadim Yanitskiy
Andreas Eversberg
Oliver Smith
Pau Espin
Harald Welte