Commit Graph

6969 Commits

Author SHA1 Message Date
Keith Whyte 991bb422d4 Implement a global switch on the network to disable call waiting.
Add a network -> callwaiting VTY command as boolean.

When this is enabled (default) there is no change to
operation previous to this commit.

When this switch is disabled with "no call-waiting" in vty
then when a call arrives, we will check if we have an active
call transaction for this subscriber, no matter if it is
establishing, established, or alerting, in any of these cases we
will return USER BUSY to the calling party.

Change-Id: I3eb6f23f7103e3002874fb5d3a30c9de952202ae
2019-09-03 11:47:13 +00:00
Neels Hofmeyr 7c8df44adc cosmetic: fix call_leg_ensure_ci() decl. arg name to match impl.
Change-Id: I576bc5c1fd63fe8048a7a6a2d06763fc3221fa49
2019-09-02 11:07:49 +02:00
Neels Hofmeyr b4ef5e7bd1 msc_a.c, CC trans: change a comment to a debug log
Change-Id: I5a3cc6219080910119b0c9ff11fc2b9eb96a06e5
2019-09-02 11:07:49 +02:00
Neels Hofmeyr 43e8d4d412 gsm48_tch_rtp_create(): check against NULL mgcp_info
osmo_mgcpc_ep_ci_get_rtp_info() might return a NULL mgcp_info, guard against
that.

Fixes: CID#203651
Change-Id: I98fe5860c49751ade1af10d99487aba259504f23
2019-09-02 11:07:49 +02:00
Neels Hofmeyr be5f033475 msc_vlr_tests: GSUP: don't care about extra IEs
To not break the msc_vlr tests by new GSUP IEs added to some of the GSUP
messages, make msc_vlr_tests only match the start of the GSUP message and not
care about extra IEs. The extra IEs are anyway seen in the expected logs.

The reason to drop the msgb_eq_data_print() is because it is useless for
mismatching lengths. It will always print only the length mismatch, instead we
need to be able to compare with what was expected.

Change-Id: I38d51eeafab04ece83e4bb87bfaa967506f97b11
2019-09-02 11:07:49 +02:00
Neels Hofmeyr 38816dab97 log, cosmetic: add "RR" to "Ciphering Mode Complete"
Distinguish the enclosed DTAP RR Ciphering Mode Complete message from the outer
BSSMAP Cipher Mode Complete message in the DEBUG log.

Change-Id: I80c69b491e2ddb932bc4295a01caaf6a903b1fe4
2019-09-02 11:06:05 +02:00
Neels Hofmeyr 43a349fd63 tweak CC cause for incoming call to unattached nr
So far we sent CC cause "Unassigned Number"

But the MSC doesn't trivially know whether the HLR has the number assigned or
not: any handset that is currently switched off would cause "Unassigned number"
to be displayed on the caller's handset.

Rather send a temporary failure cause code.

Send this cause code for all cases, because claiming that an assigned number is
unassigned is worse than rejecting an unassigned number with a temporary
failure.

Change-Id: Ia3d4f67b53fcc2654ff048fbc338e92cb763a095
2019-08-29 05:42:02 +00:00
Neels Hofmeyr 364f927ac4 vlr_lu_fsm: ignore ID_IMEISV during VLR_ULA_S_WAIT_HLR_UPD
Change-Id: I2ea4f46efa013671d93892cb07bf830393289150
2019-08-29 05:42:02 +00:00
Neels Hofmeyr 2e8f881f72 fix segfault: don't send CC REL on NULL msc_a
Apparently, if a conn disappears during an ongoing call, the CC code tried to
send a CC REL on a NULL msc_a during cleanup, which lead to a crash
(cccamp2019). Guard against that.

Crash:

 #0  msc_a_tx_dtap_to_i (msc_a=0x0, dtap=0x55a4bf2fa0f0) at ../../../../src/osmo-msc/src/libmsc/msc_a.c:1565
 #1  0x000055a4be1bb03c in trans_tx_gsm48 (trans=0x55a4bf2d52a0, trans=0x55a4bf2d52a0, trans=0x55a4bf2d52a0, msg=<optimized out>)
     at ../../../../src/osmo-msc/src/libmsc/gsm_04_08_cc.c:82
 #2  gsm48_cc_tx_release (trans=trans@entry=0x55a4bf2d52a0, arg=arg@entry=0x7ffdd731a0e0) at ../../../../src/osmo-msc/src/libmsc/gsm_04_08_cc.c:1101
 #3  0x000055a4be1bee65 in _gsm48_cc_trans_free (trans=trans@entry=0x55a4bf2d52a0) at ../../../../src/osmo-msc/src/libmsc/gsm_04_08_cc.c:278
 #4  0x000055a4be1ab654 in trans_free (trans=trans@entry=0x55a4bf2d52a0) at ../../../../src/osmo-msc/src/libmsc/transaction.c:170
 #5  0x000055a4be1bd091 in mncc_tx_to_gsm_cc (net=<optimized out>, msg=msg@entry=0x55a4bf2d3b68)
     at ../../../../src/osmo-msc/src/libmsc/gsm_04_08_cc.c:1971
 #6  0x000055a4be1bf1e5 in mncc_tx_to_cc (net=<optimized out>, arg=arg@entry=0x55a4bf2d3b68)
     at ../../../../src/osmo-msc/src/libmsc/gsm_04_08_cc.c:2049
 #7  0x000055a4be18ed63 in mncc_sock_read (bfd=0x55a4bf2563b8, bfd=0x55a4bf2563b8) at ../../../../src/osmo-msc/src/libmsc/mncc_sock.c:121
 #8  mncc_sock_cb (bfd=0x55a4bf2563b8, flags=1) at ../../../../src/osmo-msc/src/libmsc/mncc_sock.c:189
 #9  0x00007fcfad607ce1 in osmo_fd_disp_fds (_eset=0x7ffdd731a9a0, _wset=0x7ffdd731a920, _rset=0x7ffdd731a8a0)
     at ../../../src/libosmocore/src/select.c:223
 #10 osmo_select_main (polling=<optimized out>) at ../../../src/libosmocore/src/select.c:263
 #11 0x000055a4be17dd56 in main (argc=3, argv=<optimized out>) at ../../../../src/osmo-msc/src/osmo-msc/msc_main.c:723

Change-Id: Ia1bb0410ad0618c182a5f6da06af342b6d483eff
2019-08-29 05:42:02 +00:00
Neels Hofmeyr 85031ee0aa cc trans: make sure bearer cap is empty
Change-Id: I147f10f9258fc8685f2f666878dd2a655b8e4583
2019-08-29 04:58:49 +02:00
Neels Hofmeyr 61ae18c956 memleak on cc setup errors
Change-Id: I3333b90064575b270627721ace7e07d085f4ad43
2019-08-29 04:58:49 +02:00
Neels Hofmeyr 5e19b9a0bf mncc: send payload type matching chosen codec
Change-Id: Id32f32d77d24b753adb96b5393c0363439e312c2
2019-08-28 21:22:45 +00:00
Alexander Couzens 4aeb4ec182 smpp_openbsc.c: check acl before deref it
All other calls check acl before deref because in a setup
with no access policy, there won't be any acl structure

Change-Id: Ibe0256535b40351594d79baa05a0147a9f89dc26
2019-08-23 23:48:51 +02:00
Philipp Maier 47cf84d8d7 msc_a: switch RAN type back to SGs when a CSFB-Call is cleared
When a CSFB call is over the MS changes back to LTE after the call is
cleared. However, at the moment the MSC does not change the
cs.attached_via_ran flag. This may cause problems with the next call. Lets
make sure that if there is an SGs association present, the ran type is
set back to SGs when the call is cleared.

Related: SYS#4624
Change-Id: I104adecb0645b81b90ee230c57bf8b463c9e7045
2019-08-19 09:00:24 +00:00
Vadim Yanitskiy ed73ae13a1 libvlr/vlr.c: cosmetic: move message_type assignment
Change-Id: Ice7f98597b54f03069375fac56fb162f2669e7f0
2019-08-18 11:11:44 +00:00
Philipp Maier 25e7ba5b77 sgs_iface: do not use SGsAP-MO-CSFB-INDICATION for CSFB return
When the VLR/MSC receives an SGsAP-MO-CSFB-INDICATION message it sets
the RAN type back to SGs. This is wrong, the message
SGsAP-MO-CSFB-INDICATION has just an informative character. It informs
the VLR that the UE has initiated an MO CSFB call (service request).

Change-Id: I625574fc42fc915ba483db3bb406922ad6df370d
Related: SYS#4624
2019-08-16 08:59:06 +00:00
Neels Hofmeyr 4dfb2babf2 add 'encryption uea 1 2' cfg / fix ttcn3 iu tests
Recently, the ability to run UTRAN without encryption was added, but the config
for it was tied to the A5 GERAN encryption configuration. This affected
osmo-msc's default behavior of Iu, breaking osmo-msc ttcn3 Iu tests: the ttcn3
test suite sets A5 to 0 (no encryption) but still expects Iu to enable air
encryption. Fix this "regression".

Add a separate vty config option for UEA encryption, even if it does not
provide full granularity to select individual UEA algorithms yet.

As a result, Iu default behavior remains to enable encryption regardless of the
A5 config. UTRAN encryption can be disabled by the new cfg option
"encryption uea 0" alone.

Even though the new vty command already allows passing various combinations of
the UEA algorithm numbers, only '0' and '1 2' are accepted as valid
combinations, to reflect current osmo-msc capabilities.

Revert most changes to the msc_vlr test suite in commit "do not force
encryption on UTRAN" (I04ecd7a3b1cc603b2e3feb630e8c7c93fc36ccd7): use new
net->iu_encryption instead of net->a5_encryption_mask.

Adjust/add to test_nodes.vty transcript tests.

Related: OS#4144
Change-Id: Ie138f2fcb105533f7bc06a6d2e6deccf6faccc5b
2019-08-13 23:52:31 +02:00
Philipp Maier cd64af7cd4 cosmetic: make function mncc_tx_to_gsm_cc static
The function mncc_tx_to_gsm_cc() is declared as non static but only used
from within gsm_04_08_cc.c. Lets declare it as static to increase
readability of the code

Change-Id: Icd02c669cfee6dd7e6b154e303cd0f4c148c83c4
2019-08-12 08:41:34 +00:00
Pau Espin 0218814c8c Bump version: 1.4.0 → 1.5.0
Change-Id: I9b79805c703f4bb860a07bf521ddbf8cb9b29d23
2019-08-08 16:01:38 +02:00
Pau Espin b0504632c5 Fix dependency version requirements
libosmocore 1.1.0 0e8df1c7e48bcae2285c7c138bd50f932049bd24  osmo_use_count()
Fixes: 7c5346cd70

libosmocore 1.2.0 cdac620579b5bf44970b5f4bb11734fdfdf5bf59 GSM23003_MSISDN_MAX_DIGITS
Fixes: 8b0737fa71

libosmo-sccp 1.1.0 21ff9ae4264de159784a59ce4421365007bde08d osmo_sccp_addr_ri_cmp()
Fixes: c4628a3ad4

osmo-mgw 1.6.0 538d2c53d90074267e7a70a90c773baa03d6ec04 mgcp_client_endpoint_fsm.h
Fixes: c4628a3ad4

libosmo-netif 0.6.0 needed by osmo-mgw 1.6.0

libsmpp34 1.14.0 required due to needed heap allocation function overrides from smpp34_heap.h

Change-Id: I1cd2c7317933b416b74a412f10e4cad16ecd6e59
2019-08-08 15:56:25 +02:00
Neels Hofmeyr 106ba523c8 vlr_lu_fsm: fix missing event for IMEISV
Event VLR_ULA_E_ID_IMEISV is listed as permitted in VLR_ULA_S_WAIT_LU_COMPL,
but is missing from the switch() on the incoming event. So, sending an IMEISV
identity during the WAIT_LU_COMPL state would crash osmo-msc.

When receiving an IMEISV, vlr_subscr_set_imeisv() in turn calls
vlr_subscr_set_imei(), so as far as the lu_fsm is concerned, receiving an
IMEISV is identical to receiving an IMEI, and it can continue to send a Check
IMEI request to the HLR. Thus simply add VLR_ULA_E_ID_IMEISV to the
VLR_ULA_E_ID_IMEI switch case.

Change-Id: I11106cb108a4b1406ff9a8b8ff5761440a274dad
2019-08-08 02:08:55 +02:00
Neels Hofmeyr 7ea0db8f29 doc/sequence_charts: fix naming of mncc_fsm to mncc_call
mncc_fsm.[hc] were renamed to mncc_call.[hc] during patch review, which failed
to carry through to this sequence chart.

Also fix the MNCC_ST_* to MNCC_CALL_ST_* and MNCC_EV_* to MNCC_CALL_EV_*.

Change-Id: I03ee1b43ab95dca3c43fdb9e92dc158aad5a4203
2019-08-05 23:25:17 +02:00
Neels Hofmeyr 62c0a2a4ab doc/sequence_charts/mncc_fsm.msc: add SIP messages, tweak
- add SIP messages, taken from OS#1683
- change some wording and clarify some message ordering
- have a separate sipcon1 and sipcon2 for the MO and MT sides

Change-Id: I6782e416dbd8ee88d093cbef722b0c5084f3865c
2019-08-05 23:25:16 +02:00
Neels Hofmeyr 14c34d8bec add msc_vlr tests for UMTS without ciphering
Following I04ecd7a3b1cc603b2e3feb630e8c7c93fc36ccd7, have tests for UMTS
authentication both for cases with and without encryption.

- Rename test_umts_authen_utran to test_umts_auth_ciph_utran() (uses
  encryption).
- Again add test_umts_authen_utran() not using encryption.
- Likewise with test_umts_authen_resync_utran().

Some permutations are still missing, like UMTS AKA on GERAN with encryption
enabled; not bothering at the moment.

Related: OS#2783
Change-Id: I54227f1f08c38c0bf69b9c48924669c4829b04b9
2019-08-05 23:24:47 +02:00
Neels Hofmeyr a4d7a76816 do not force encryption on UTRAN
Remove the conditions that always enable encryption on UTRAN.

We so far lack an explicit configuration for UTRAN encryption, and this patch
does not add any either. Instead, whether UTRAN encryption is enabled is simply
triggered on whether GERAN has A5 encryption enabled (A5/n with n > 0). Though
GERAN and UTRAN encryption are not technically related at all, this makes UTRAN
behave like GERAN for now, until we implement a proper separate configuration
for UTRAN encryption.

Adjust the msc_vlr_test_* configuration by setting the net->a5_encryption_mask
such that the expected output remains unchanged. A subsequent patch
(I54227f1f08c38c0bf69b9c48924669c4829b04b9) will add more tests, particularly
cases of UTRAN without encryption.

Adjust manual and vty doc.

Related: OS#2783
Change-Id: I04ecd7a3b1cc603b2e3feb630e8c7c93fc36ccd7
2019-08-05 23:24:47 +02:00
Neels Hofmeyr 75bdbbf45d manual: adjust and fix auth and ciph docs
Change-Id: Iffe159d4c0e0e9439f8719e0ddd28f06d4c80d9f
2019-08-05 23:24:47 +02:00
Pau Espin 3e1e47679d Remove undefined param passed to logging_vty_add_cmds
Since March 15th 2017, libosmocore API logging_vty_add_cmds() had its
parameter removed (c65c5b4ea075ef6cef11fff9442ae0b15c1d6af7). However,
definition in C file doesn't contain "(void)", which means number of
parameters is undefined and thus compiler doesn't complain. Let's remove
parameters from all callers before enforcing "(void)" on it.

Change-Id: Ia2b24ffd7f9cbb271fcdb979b851f3a07b9d6d3e
Related: OS#4138
2019-08-05 16:07:08 +02:00
Keith Whyte ff17f8fd8a Set coding in mncc_set_cause()
GSM 04.08 10.5.4.11

The Release indication needs to have the Coding Standard set.

For phones that would display a message on screen, such as
"Number not in use", if the coding standard is not defined,
the display may show "Error in Connection"

Change-Id: Ib28b62a41d433e231cff5910d19455296b284df6
2019-08-02 08:08:33 +00:00
Neels Hofmeyr 4014e481d4 minor comments in msc_vty.c
Change-Id: I83d8c778190adb1e74debc8f8ddac6996de7c513
2019-07-29 18:39:50 +02:00
Pau Espin 637fc0218b doc: Add Osmux documentation to User Manual
Depends: osmo-gsm-manuals.git f3a734e6777a902abfb03257277454c7a879aeb7
Change-Id: I70c488c3d9b05599b834a8608e6361c8aa43ef31
2019-07-25 10:39:07 +00:00
Oliver Smith c0a5e71d0e vlr_lu_fsm.c: don't send LU reject twice
Don't call tx_lu_rej() in the "vlr_lu_compl" FSM. It is always getting
called in the parent "lu" FSM and is therefore redundant:

_vlr_lu_compl_fsm_done(fi, VLR_FSM_RESULT_FAILURE, cause)
-> osmo_fsm_inst_state_chg(fi, LU_COMPL_VLR_S_DONE, 0, 0)
-> vlr_lu_compl_fsm_dispatch_result()
-> lu_fsm_wait_lu_compl()/lu_fsm_wait_lu_compl_standalone()
-> lu_fsm_failure()
-> lfp->vlr->ops.tx_lu_rej()

I have noticed the bug with the TTCN3 tests. This patch fixes
TC_lu_imsi_auth_tmsi_check_imei_{nack,err} after stricter checking
in [1] and also TC_iu_mo_crcx_ran_reject.

[1] I836f76242463789c4c003feec757714827f2a31b (osmo-ttcn3-hacks)
Change-Id: I127b27937613ea0ff29d67991c0414fca6d441d9
2019-07-24 09:08:22 +02:00
Eric Wild fdda613af5 turn -Werror=null-dereference into a warning
There is unfortunately no way to suppres this witha pragma,
and gcc 9 uncovers quite a few new instaces with enabled LTO that can't/won't be fixed

Related: OS#4123
Change-Id: I615bb5be3671022c6b821575a61f945b50e8f2a5
2019-07-22 19:56:56 +00:00
Alexander Couzens efa7b97133 replace osmo_counter with stat_items
osmo_counter will be soon deprecated. Use the newer and more flexible
osmo_stat_item instead.

Depends on: Id2462c4866bd22bc2338c9c8f69b775f88ae7511 (libosmocore)
Change-Id: I6a20123b263f4f808153794ee8a735092deb399e
2019-07-18 14:50:51 +00:00
Thorsten Alteholz 3a357de631 fix spelling detected by lintian
Change-Id: I01e54b5cf111677079a8ad57645d3ceb7834702a
2019-07-16 20:12:03 +00:00
Oliver Smith aa0a295a3d contrib/jenkins.sh: run "make maintainer-clean"
Related: OS#3047
Change-Id: Ifc91733c7fe0c5a002037523116014bf270997ef
2019-07-11 03:42:36 +00:00
Vadim Yanitskiy 09ce3dcbdf libmsc/msc_vty.c: print subscriber expiration time
Change-Id: I092691a8c443f4c5ed4d33de2e551fef592c1baf
2019-07-09 07:14:09 +00:00
Vadim Yanitskiy d14422a236 Fix: add missing semicolons to OSMO_ASSERT statements
Change-Id: I4fae5fbab5fdbcce35906601d4f1031d971f4931
2019-07-09 07:13:45 +00:00
Vadim Yanitskiy 1c50044efb libmsc/ran_msg_iu.c: fix: properly handle SAPI IE of RANAP_DirectTransfer
The RANAP DirectTransfer message may contain an optional SAPI IE.
Thanks to our TTCN-3 tests (and Wireshark!), it was discovered
that this IE is ignored, so even if the MO SMS related messages
arrive on SAPI 3 (as per GSM TS 04.11, section 2.3) OsmoMSC sends
MT messages on SAPI 0.

In ran_iu_decode_l3() we need to check if the SAPI IE is present,
and tag the NAS PDU message buffer with a proper DLCI value.

This change makes the failing SMS related test cases pass.

Change-Id: I728b55b04e87fc23be6d4f8735e8cad82b6f640e
2019-06-20 16:54:45 +00:00
Vadim Yanitskiy 7659482318 libmsc/gsm_04_11.c: do not abuse LOG_TRANS() in gsm411_alloc_mt_trans()
This change is similar to I6b68a0f0b32eb126e0f7e914a314130254d28467.
If we 100% sure that trans == NULL, it makes more sense to use
generic LOGP(DLSMS, LOGL_*, ...) call, so the logs can reflect
more information than such dummy prefix:

  trans(NULL NULL callref-0x0 tid-0) ...

Change-Id: I3c1e633aee5dd7cd0d367404a3def9cffe0b3baa
2019-06-20 16:54:45 +00:00
Vadim Yanitskiy 59e0c6b46a gsm_04_11_gsup.c: fix broken reference counting for vsub
This change is similar to I5540556b1c75f6873883e46b78656f31fc1ef186.
In gsm411_gsup_rx() we do call vlr_subscr_find_by_imsi(), which
increases subscriber's reference count by one using the function
name as the token. However, we never release this token, so the
reference count grows on every received GSUP FORWARD-SM message.

Change-Id: Ic729beb5f94cbbfbb251bc9ab66a5e7b799286c0
2019-06-20 16:54:45 +00:00
Pau Espin fe5b7046ea sms_queue.c: Improve misleading log line
Otherwise when read in a log file it seems it's really going to send 20
sms even if there's none to send.

Change-Id: Ieb9bb61a90f295d2ba5fb67a2abee2d30785876d
2019-06-20 10:45:37 +02:00
Vadim Yanitskiy 718f32fa88 libvlr/vlr.c: do not expire subscribers if periodic LU is disabled
When periodic Location Update is disabled (T3212 = 0), it was noticed
that OsmoMSC does expire subscribers quite soon - after 60 seconds
(VLR_SUBSCRIBER_LU_EXPIRATION_INTERVAL) since the last LU.

In order to avoid that, we need to check T3212 timer value in
vlr_subscr_expire_lu(), and if it's equal to 0, do not expire
anybody until the explicit IMSI Detach.

Change-Id: I2ead2241a3394dbdd5417f4554190df3fd698af2
2019-06-19 14:36:28 +07:00
Vadim Yanitskiy 188dd5f4df tests/msc_vlr: fix: do not pass RAT type to expect_bssap_clear()
The function name implies OSMO_RAT_GERAN_A, and it has nothing
to do with other OSMO_RAT_* types. Found using clang:

  warning: too many arguments in call to 'expect_bssap_clear'
           expect_bssap_clear(OSMO_RAT_GERAN_A);
	                      ^^^^^^^^^^^^^^^^

Change-Id: Id3a3af33fcc5da4ca4c48a2f589a69f3568d2586
2019-06-19 02:05:08 +07:00
Daniel Willmann 7a2c9c91e9 manuals: Include overview chapter about counters
Change-Id: I8c3e8bcda27f35118ab0e3d75621a02eec86f15c
2019-06-18 18:07:43 +02:00
Daniel Willmann fb73e0c654 manuals: Update counter/vty documentation
Change-Id: I1ef0e8ae166d7fdc5e85716a961e8387d26bdd2c
Related: OS#1700
2019-06-18 12:51:07 +00:00
Daniel Willmann a794ff4f46 manuals: Add script to update vty/counter documentation from docker
Change-Id: I3f5573f81460b40d4606fbcf0febcfd078a7bdca
Related: OS#1700
2019-06-18 12:51:07 +00:00
Vadim Yanitskiy b3bdc768be libmsc/gsm_09_11.c: do not suppress rc of gsup_client_mux_tx()
Change-Id: Ide2440188fb6fe1c54681fef8ec4fed9e6da66e2
2019-06-17 22:23:09 +07:00
Vadim Yanitskiy 6632251572 libmsc/gsm_09_11.c: do not abuse LOG_TRANS() in gsm0911_rcv_nc_ss()
If we 100% sure that trans == NULL, it makes more sense to use
generic LOGP(DSS, LOGL_*, ...) call, so the logs can reflect
more information than such dummy prefix:

  trans(NULL NULL callref-0x0 tid-0) ...

Change-Id: I6b68a0f0b32eb126e0f7e914a314130254d28467
2019-06-17 22:13:03 +07:00
Vadim Yanitskiy e0da446e76 libmsc/gsm_09_11.c: fix broken reference counting for vsub
In gsm0911_gsup_rx() we do call vlr_subscr_find_by_imsi(), which
increases subscriber's reference count by one using the function
name as the token. However, we never release this token, so the
reference count grows on every received GSUP PROC-SS message.

Change-Id: I5540556b1c75f6873883e46b78656f31fc1ef186
2019-06-17 21:57:30 +07:00
Vadim Yanitskiy 944d6a2acb libmsc/gsm_09_11.c: avoid double zero-initialization of gsup_msg
Change-Id: Ib991b01534499401e7a0c3de49ceba770fdd9b48
2019-06-17 21:57:30 +07:00