Commit Graph

7017 Commits

Author SHA1 Message Date
Alexander Couzens 2aaff7513b Validate the choosen UTRAN encryption algorithm
RANAP Security Command can include an encryption IE. If it includes
it the RNC can still ignore it (e.g. unsupported encryption) and
return the Security Command Complete with an choosen encryption IE:
"no encryption".
Validate the encryption element and ensure the encryption is included in
the encryption mask.

Closes: OS#4144
Change-Id: Icfc135c8b8ae862defe7114db492af600c26407f
2021-11-05 01:29:05 +01:00
Harald Welte 505a94a610 Make UTRAN encryption algorithms configurable
Allow the user fine-grained control over which UMTS encryption
algorithms are permitted, rather than always permitting UEA1 and UEA2
or neither.

This brings the handling of UEA in line with the handling of A5 for

Change-Id: I91f9e50f9c1439aa19528f887b83ae9de628fcfd
Closes: OS#4144
Depends: osmo-iuh.git I6d2d033b0427bdc84fee61e0f3cb7b29935214bf
2021-11-05 01:29:05 +01:00
Harald Welte 274b70f1a8 Fix enabling of UMTS UEA encryption
The existing code allowed the user to configure UMTS encryption in the
vty, but we never actually passed this information down to RANAP.  As a
result, the RAN had no chance of ever enabling encryption on the air

Change-Id: Ieaaa6b23b7337b7edb902fad8031e195e0c5e9d2
Related: OS#4144
2021-11-05 01:29:05 +01:00
Oliver Smith efa6c5b7d6 Revert "Turn some compiler warnings into errors"
Do not turn some compiler warnings into errors by default. This patch
was added before --enable-werror was available.

We build with --enable-werror during development and in CI. If the code
is built with a different compiler that throws additional warnings, it
should not stop the build.

This reverts commit 34f012639d.

Related: OS#5289
Change-Id: Ideff462157a034e053e5e7049605dd8d24440905
2021-11-04 13:03:20 +00:00
Vadim Yanitskiy c6921e5068 mncc: rework passing GCR over the MNCC interface
Using *unpacked* 'struct osmo_gcr_parsed' in the MNCC PDUs makes
the protocol even more complicated than it currently is, and
moreover complicates implementing MNCCv8 in the ttcn3-sip-test.

Replace 'struct osmo_gcr_parsed' in 'struct gsm_mncc' with a
fixed-length buffer, which is supposed to hold the Global Call
Reference encoded as per 3GPP TS 29.205.

Indicate presence of GCR using the MNCC_F_GCR flag.

Change-Id: I259b6d7e4cbe26159b9b496356fc7c1c27d54521
Fixes: I705c860e51637b4537cad65a330ecbaaca96dd5b
Related: OS#5164, OS#5282
2021-11-03 21:12:52 +00:00
Vadim Yanitskiy e0661105ac smpp_smsc: use osmo_talloc_replace_string() in smpp_smsc_conf()
Change-Id: Ie3dc0d9673a0410a5908cba77fc352d581bac6f6
Fixes: CID#240711
2021-10-30 02:27:21 +00:00
Vadim Yanitskiy 97b8e76b50 libmsc: struct smsc: drop 'const' qualifier from bind_addr
This pointer holds an address of a string allocated on heap.

Change-Id: I50145b57494aaccd4793f7b0ed2f65a6641db929
Related: CID#240711
2021-10-26 18:42:57 +00:00
Vadim Yanitskiy 583883bf2e libmsc: fix NULL pointer dereference in trans_lcls_compose()
Change-Id: I96342de4cb3e0bc3d3cac7538d4517aa211e8a38
Fixes: CID#240768
2021-10-26 10:25:39 +03:00
Keith Whyte a1a70be593 Add support for LCLS to the MSC
This commit is largely based on work by
Max <>

Adds LCLS parameters for A-interface transactions
This commit also adds a vty option to facilitate globally
disabling LCLS for all calls on this MSC.

Add a global call reference (GCR) to MNCC and therefore
bump the MNCC version to version 8. (This commit has to be
merged at the same time as the corresponing commit in the
osmo-sip-connector for mncc-external use.)

Depends: osmo-sip-connector Id40d7e0fed9356f801b3627c118150055e7232b1
Change-Id: I705c860e51637b4537cad65a330ecbaaca96dd5b
2021-10-25 10:07:48 +00:00
Harald Welte 890ece1277 smpp_mirror: Don't allocate msgb's for unrealistic amounts of memory
If the remote ESME would send us 0xffffffff as length field, don't try
to allocte 4GB of memory, but bail out.

Change-Id: I561f75210811826de06ea1673eca1df24faaa210
Fixes: CID#240738
2021-10-25 09:58:51 +00:00
Harald Welte 065b23ae5b smpp_mirror: Factor-out reset of SMPP read state
Change-Id: Ib9f210316eaddae19cb74db92b922016864dd786
2021-10-25 08:20:00 +02:00
Harald Welte 30811f865c smpp: Fix help string in vty for "alert-notifications"
Without the 'no' prefix, this command is about enabling, not disabling

Change-Id: I3019a0d9271e73030b9d7cd555761e17358ec812
2021-09-28 21:40:40 +02:00
Oliver Smith 4ac59538e9 contrib/ require libosmo-ranap >= 0.7.0
Sync with, debian/control. Nowadays
complains about this.

Related: I8dbbd63fdbd35f5f90f20dcd7e710078f2d9e12a
Change-Id: Ib3b9d6af8f9d4e807ff4106f95f0a5b13834a126
2021-09-16 14:34:18 +02:00
Philipp Maier ce1f973283 msc_vlr_test: remove DLMGCP log messages from unit test output
The log output of libosmo-mgcp-client has changed. This change causes
the unit tests to fail because the log output does not match anymore.
Lets disable the DLMGCP log output since it is of minor importance
for VLR testing anyway.

Change-Id: Id197e4ab9ba12e284299ef520edee9c362513bf1
Related: SYS#5091
2021-09-14 12:06:45 +02:00
Pau Espin 35ff852379 cosmetic: Fix typo in comment
Change-Id: Ic167bb4e12c657f917ad4f76ef65f46753c5f7a5
2021-08-25 12:40:36 +02:00
Pau Espin 017ba1fa30 vlr_sgs: Balance use_count incremented in vlr_sgs_loc_update
During a recent pcap trace, it was spotted that subscriber coming from
SGs had a use count with 16 "SGs" items, and later it incremented to 17.
Further investigation shows that the related use_count item was never
decreased, meaning every time an SGs-LU was sent by the MME, the item
was incremented further and never decremented.

Let's rename the item to be referenced while in LU, and then decremented
when LU is done. At that time, either the LU was accepted and the
subscriber object has a use_count item "attached", or it was rejected
and we already sent the reject messages, so we are fine deleting it if

Related: SYS#5337
Change-Id: I22c386f02ffa57428f700b003cc2cf23133598d0
2021-08-24 14:59:27 +02:00
Pau Espin 25b3ffb5f4 vlr_sgs.c: Fix missing use_count decrease in vlr_sgs_imsi_detach
it was recently observed in a pcap trace with gsmtap_log that the
use_count contained a "vlr_sgs_imsi_detach" item despite no related
message was seen near by. Further investigation shows that there's an
unbalanced get+put code path, introduced by an early return added to fix
another issue.

related: SYS#5337
Fixes: 0803d88d9a
Change-Id: I91ae956e50fca2f4d0e1d145d60ccb0ebfb409e9
2021-08-24 14:59:12 +02:00
Pau Espin eb1b3ab6da vlr_auth_fsm.c: Simplify function auth_fsm_wait_ai_resync
Change-Id: I2070002dea18b728deef5547b4deb6166cfcab6a
2021-08-24 12:25:40 +02:00
Pau Espin 13f17c13c3 vlr_sgs.h: Set proper logic order of items in enum sgs_ue_fsm_state
Set order of states in the same order as they appear in the specs (see
chapter 4.2.2 mentioned above the enum).
Furthermore, from FSM state transition point of view it also makes sense
to put them in this new order, since one should pass through

Change-Id: Ia9216965e9f30caedffa3cb53d14da7f7fd37b4e
2021-08-24 11:49:05 +02:00
Pau Espin 25b6847f46 cosmetic: fix typos in comments
Change-Id: I89e8eba379f83bcf3061601b02af9f10dbca0f22
2021-08-24 11:42:15 +02:00
Philipp Maier 945b0606cf running.adoc: add section about MGCP configuration
The manual seems to lack a section about how the MGW is set up. In the
osmo-bsc manual we have a "Configure MGCP to connect to an MGW" section
under the "Configure primary links" section. We should have the same
thing in the osmo-msc manual as well.

Change-Id: I5501739e63860c436ff606bc2758b495258cd2b9
Depends: osmo-mgw I47e7ff858d5067b46d52329be5f362ff61c0dff8
2021-08-03 11:36:25 +02:00
Neels Hofmeyr ae98b97382 implement CM Re-Establish for voice calls
Related: SYS#5130
Change-Id: I6fa37d6ca9fcb1637742b40e37b68d67664c9b60
2021-07-29 22:40:59 +02:00
Neels Hofmeyr 5bdba0d48d add vlr_subscr_find_by_mi
Will be used by I6fa37d6ca9fcb1637742b40e37b68d67664c9b60
"implement CM Re-Establish for voice calls"

Related: SYS#5130
Change-Id: I5291d098a02268bd1c2e30195ae61e4a13e8709c
2021-07-28 18:37:48 +02:00
Neels Hofmeyr 0c076bc75a add sequence_charts/call_reestablishment.msc
Change-Id: I6ebc3a0555ef82e1c027ca75b45f8147a3e07ee6
2021-07-28 18:37:48 +02:00
Pau Espin e53ecde83e msc_a.c: Allow MSC_A_EV_CN_CLOSE in state MSC_A_ST_RELEASING
As seen in a running osmo-msc:
Event MSC_A_EV_CN_CLOSE not permitted"

Also seen in several unit tests, which need update.

The action event handler for that state is actually already
expecting/handling the event by ignoring it, so we should allow it.

Change-Id: I4d30cffab693529aab3ba736419dec116a4dd7ef
2021-07-12 16:37:24 +02:00
Pau Espin 42ac3665c8 sgs_iface.c: Improve logging when paging over SGs
Change-Id: I745d3e904583ddf170ec1a1ceac5a49b72e813e3
2021-07-12 16:37:19 +02:00
Pau Espin 1086e20754 vlr_sgs: Drop recorded LastEutranPlmnId when UE no longer associated
Change-Id: I0dcb1850ab4f6f7d42bfcb19665ddeae2d3b057a
2021-07-02 21:46:49 +02:00
Pau Espin 8252959b74 sgs: Use available API to set vlr subscr LastUsedEutranPLMNId
This way it's easier to find out where the vlr subscriber field is set.

Change-Id: I203de82accc896c196cb70037af89be6dc826c05
2021-07-01 18:14:39 +02:00
Neels Hofmeyr 73d093a806 improve logging on encryption, for Ciphering and HO
log the algorithm and all keys instead of just Kc.

Change-Id: If7871dedda2b1fb115e6f00da1379ff2e3d68a10
2021-06-24 01:44:20 +02:00
Neels Hofmeyr db07fdcbfc support A5/4 in inter-BSC handover
Forward the Kc128 key to the new BSS in BSSMAP Handover Request.

Depends: Ieb6e43eef9e57281d54d4b7c63664668df5aef3e (libosmocore)
Change-Id: Id5ce995a741c8e469a50a0c46e53c06a2378bb7e
2021-06-24 01:44:17 +02:00
Neels Hofmeyr 6ce2edcac1 support A5/4 in Cipher Mode Command
Related: SYS#5324
Change-Id: I780a739b9bfbefd4f58be051794fe1a491823e67
2021-06-24 01:27:47 +02:00
Eric Wild a7f8020bfa vty: allow A5/4 encryption in config
Add A5/4 to the internal mask of allowed algorithms.

(Not actually working yet, A5/4 implementation follows in other

Related: SYS#5324
Change-Id: I5b46aaa8579f8d069ca39caf996a8795ffe63dd7
2021-06-24 01:27:47 +02:00
Neels Hofmeyr cdcfc80176 ran_msg_a.c: use gsm0808_create_cipher2()
Use new API in Cipher Mode Command to prepare for A5/4 support.

Depends: Ib3906085e0c6e5a496a9f755f0f786238a86ca34 (libosmocore)
Related: SYS#5324
Change-Id: Ib238d367b8d5d07b6ab4cb2e48fbf4ce22ca4476
2021-06-23 23:58:59 +02:00
Neels Hofmeyr 07c8b7cb2e osmo-msc main: use osmo_select_shutdown()
Change-Id: I75f4637c051ed44628e65dab1bdbbf28dcc9626f
2021-06-22 23:34:34 +02:00
Vadim Yanitskiy 435f67f43c msc_tx_common_id(): fix potential NULL pointer dereference
Reported by GCC 11.1.0.  msc_a_vsub() may return NULL.

Change-Id: Iebdd6399e819a03258398e6b7b453bda37e45a20
2021-06-06 15:50:42 +02:00
Pau Espin 2e21a68e83 Use new osmo stat items/ctr APIs
Generated using several semantinc patches with spatch.

Change-Id: I3ee853539949a763a309856bf2e7196415b23741
2021-06-04 16:53:16 +02:00
Pau Espin 6710670cb1 Fill Last Used E-UTRAN PLMN Id when in CSFB
Since recently, osmo-bsc behaves strictly as per specs, meaning it will
only send the "Cell selection indicator after release of all TCH and SDCCH IE"
in RR Channel Release iff:
* "Last Used E-UTRAN PLMN Id" was received in the CommonID sent MSC->BSC
* "Last Used E-UTRAN PLMN Id" was received insider "old BSS to new BSS Information"
  in the HandoverRequest sent MSC->BSC.
On the other hand, CSFB_Indicator from ClearCommand MSC->BSC is nw
ignored and not taken into account.

Hence, let's update osmo-msc to also behave correctly by sending the
Last Used E-UTRAN PLMN ID at CommonID tx time to avoid regressions in
CSFB support when running against newer osmo-bsc.

Let's keep sending the CSFB Indicator in ClearCommand as we used too, in
order to keep compatibility with older BSCs (as per spec).

Related: SYS#5337
Change-Id: Ic5f175b179973d0a50d94f00e15f5a3e332605fc
2021-04-29 14:22:33 +02:00
Vadim Yanitskiy a361cab54a SMS-over-GSUP: notify sender if no transaction found
Change-Id: I4f564610fadbfdbbc33de267786534a5405319f6
2021-04-04 14:09:32 +00:00
Oliver Smith 5a44e5f46e contrib/ depend on dbd-sqlite3
Add the missing runtime dependency to the sqlite3 driver of libdbd.
The library does not provide a pkgconfig file, so using "pkgconfig(...)"
as done in the BuildRequires is not possible. Write both the OpenSUSE
and CentOS name with an if..else.

  <0009> db.c:648 Failed to create database connection to sqlite3 db 'sms.db';
                  Is the sqlite3 database driver for libdbi installed on this system?

Change-Id: Ia972944c300aecbb6ec460b2362aabff459baefd
2021-03-09 14:48:59 +01:00
Pau Espin 13c2153c94 Depend on newer libosmo-ranap
I forgot to ammend the release previous release commit with this.

Change-Id: I8dbbd63fdbd35f5f90f20dcd7e710078f2d9e12a
2021-02-23 20:47:57 +01:00
Pau Espin 59764b6f3e Bump version: → 1.7.0
Change-Id: I1bf25c0c01efd6f93df8ac2b26868dd1579b3531
2021-02-23 20:22:34 +01:00
Pau Espin cad22fde22 tests: Replace deprecated API log_set_print_filename
Change-Id: I4c1116aafb155b5fb216b01054332dcb91c18401
2021-02-19 14:11:26 +01:00
Pau Espin 8d02d99158 tests: Explicitly drop category from log
Let's disable category here since we don't care about its formatting here.

In any case, every test relying on logging output validation should
always explicitly state the config to avoid issues in the future if
default values change.

Change-Id: I089c0001fc75e81558c3e860827e4d434cf1eab3
Related: OS#5034
2021-02-19 14:10:50 +01:00
Neels Hofmeyr ac6ccd9295 gsm_network: drop unused neighbor_list member
Change-Id: I57463ab94773de1377526f9a1c72fe3d37a31827
2021-02-18 14:16:18 +01:00
Vadim Yanitskiy dd466cfa59 fix _gsm48_cc_trans_free(): send MNCC REL.ind on Clear Request
Calling gsm48_cc_tx_release() before mncc_release_ind() has a side
effect: the former may change CC state to GSM_CSTATE_RELEASE_REQ.
This makes the later send MNCC_REL_CNF instead of MNCC_REL_IND, so
if one of the call leg disconnects due to RF failure, the other one
will not be terminated correctly.

Makes both TC_{mo,mt}_call_clear_request TTCN-3 test cases pass.

Change-Id: I3ad4a99757878de3796027325627c87d9a4e93f1
Related: Id16969fe0de04445d1320a96d35cf1d48cc8cf09
Related: SYS#5340
2021-02-09 18:34:40 +01:00
Neels Hofmeyr dbb3c723a4 msc_vlr_tests: make independent of libosmocore talloc
The msc_vlr_tests verify whether any of the tests run contain msgb or
talloc memory leaks. So far they did so by fixating a specific number of
talloc blocks, which may break by library implementations changing.

Instead, verify that the test leaks no allocations by comparing talloc
blocks before and after each test.

When a leak is detected, print the full talloc report to stderr, which
makes the expected output mismatch the actual output and fails the test.

Related: OS#4311
Change-Id: I8537fa76d460c951302932a1bad4299f7fe398c9
2021-02-07 23:59:59 +01:00
Vadim Yanitskiy 33144f146a ran_a_decode_cipher_mode_reject(): use gsm0808_get_cause()
Change-Id: I56dd920b931e769ba4d268b09700fe3c9fca4fc6
2021-02-05 20:16:52 +01:00
Vadim Yanitskiy 566ce11df5 submit_to_sms(): fix the use of deprecated gsm_septets2octets()
Change-Id: I9985972f0b1d2b71bfd133c5004201a3a0ffcbd0
2021-02-05 20:15:08 +01:00
Pau Espin 46de4aafa6 .gitignore: Ignore new autofoo tmp files
Change-Id: Ifb9dfe84575451fdfc9d059bab428da82b448107
2021-02-04 12:51:41 +01:00
Oliver Smith 09dac13c4d set -std=gnu11
Change-Id: Ie93915f06a8907a80b4f081a00c1cde395015595
2021-01-28 09:28:26 +00:00