In TC_pdp4_act_deact_with_single_dns we activate, deactivate and then
re-activate a PDP context. Hoewver, we re-use the same variable and
don't reset the state in between. This results in the second PDP CTX
activation to include an end-user-address (static IP allocation), which
OsmoGGSN doesn't implement.
Before osmo-ggsn Change-Id Iac8868438655fe4e5e07d167d7dbd6273dbb7678,
the test passed as osmo-ggsn simply ignored the requested static address.
After that change, we reject static addresses and hence the test starts
to fail.
Change-Id: I1b1869bc2cee39c8fddd8fa63f48bdaa6a65e462
Related: OS#5097
This adds IUT fsm state instrospection via the CTRL interface.
docker-playground will need to set "mp_gbproxy_ip" in its configs.
Change-Id: I272e43b9be8ba53d8a815e8ab099c939f63413a7
The expectations of this test case were wrong. The IUT would first
accept() an additional connection and then close() it immediately.
Since there may be other messages, like TIME.ind and DATA.ind, the
'alt' statement would not match successful connection result, and
instead would unblock the flow due to timeout.
The titan.TestPorts.UNIX_DOMAIN_SOCKETasp had to be changed [1] to
send UD_connect_result with ERROR if recv() returns zero or a negative.
[1] https://github.com/eclipse/titan.TestPorts.UNIX_DOMAIN_SOCKETasp/pull/4
Change-Id: I898b8b14515d79766b12d652ebb1ddf834e2863c
Those tests predated SGSN pooling support and needed to update
their assumptions to also cover other SGSNs in the pool.
We could come up with more complex, auto-scaling mechanisms to handle
different numbers of SGSNs, but that would make the code quite
unreadable, as it means we cannot use interleave but have to dynamically
activate altsteps for counting. I went for the simpler approach
instead.
Change-Id: I8c3ec4e00b78cff40ac72bc857161fa7abc352b0
Related: OS#4968
We used to have no distinction between endpoints and NSVCs, meaning
that we could not have more than one NSVC per endpoint, which in turn
meant it was ok to iterate the list of NSVCs for generating the
endpoint lists in the SNS-CONFIG payload.
With Change-Id I05a50b966b8ce93497372ca403d40fd383dd35f7 we remove that
constraint and introduce an actual local IP endpoint list. Let's use
that one for SNS-CONFIG.
Change-Id: Ifa91510430a017fa29592a3d5fa2a3697d29c9da
Thew newly-introduced "endpoint" mode of NS_Provider_IPL4 allows us
to have multiple NS-VC within one endpoint. NS_CT takes care of
creating the NS_Provider_IPL4 components on-the-fly, and then
associating the per-NSVC components with it.
For the user nothing changes, other than that he can now configure
multiple NSVCs with identical local IP + port, which was not possible
before.
Change-Id: Id7360f17e528706e8145d33a14550789d50cded9
This adds a new, optional "endpoint" mode to the IPL4 provider,
where we remove the constraint of "one NSVC per provider", which
in turn is a pre-requirement for real IP-SNS support with full
mesh of NS-VCs.
This code doesn't introduce any users of this "endpoint" mode yet.
Change-Id: I05a50b966b8ce93497372ca403d40fd383dd35f7
There's some offset between Tx and Rx path, so we need to account for
differences counting and finding out USF blocks didn't arrive.
Change-Id: I868e7d24c8bdc9b85797f8fe4f9ee1bc5a3d1adb
Also change a bit expectations, since it can actually happen that DL
blocks for GPRS-only MS never signal USF for itself, which is
still fine.
Change-Id: Iedff87cedf55ab18b32bd0f159d1145901878203
This commit fixes a regression introduced in [1]
Change-Id: I107039f1ff44ae8c41d690f5f293ed136c17586b
Fixes: [1] Ia9f366ca1fdad700a90ca3367e43523f7bac39a1
The PCUIF connection involves a lot of frequent messages, such as
the TIME.ind and since recently DATA.ind with len=0. As a result,
the test suite logs are getting unreadable due to lots of coding
warnings and port queueing notifications.
This change is aimed to improve the situation a bit, by establishing
the PCUIF connection only for those test cases which actually use it.
Side effects:
* TC_pcu_socket_verify_info_ind becomes reliable, because the
PCUIF establishment is done after the RSL bootstrapping;
* TC_pcu_socket_connect_multi starts to fail, because it used
to pass due to timeout, since not all messages are handled
in the 'alt' statement.
Change-Id: I09ccb65ce94a41ffdad4e93da650c3f32d422af4
Related: OS#5083
PCUIFv9 compatibility has been dropped in [1], so now we can tell
TITAN's RAW codec what kind of padding to expect in received
messages and to append to encoded messages. This eliminates
thousands of warnings about unhandled tail octets.
Change-Id: Id5ab587f7984a91f3603efc5d427aa4cb4fa6757
Related: [1] Ia9f366ca1fdad700a90ca3367e43523f7bac39a1
Related: OS#5083
In f_routing_area_update() we are sending a RAU Complete to the SGSN
and then immediately afterwards send a GTP-U from the simulated
GGSN to the SGSN. That GTP might reach the SGSN faster than the
RAU Complete, resulting in a test failure.
Change-Id: Ic489e0857115cf24965e413a39918edc5a8f44f8
In recent osmo-pcu commits, initial fn was changed to invalid value -1,
in order to be able to detect FN jumps. previously, the initial value
was set randomly to 0, which was wrong anyway because first FN received
from the BTS could be any other FN counted by the BTS at that time.
This makes some tests fail because they send RACH.ind + RTS.ind to
receive the Imm Assignment, and the Request reference in the Imm Assign
was calculated on the invalid unset FN "-1", hence it won't match test
expectancies.
In order to fix it, simply make sure the TDMA clock is initiated by
sending a DATA.ind to the PCU before tests start doing stuff
(f_init_raw() is blocked waiting for BTS_EV_SI13_NEGO).
Related: osmo-pcu.git Change-Id I29fb27981597edc69abb976049ba41aa840488cb
Related: OS#5020
Change-Id: I00c4dd9133ec9a236bf28fb8cb0afd0615791012
PCUIF will be updated to always send DATA.ind for each expected block FN
on any activated PDCH slot, irrespective of whether valid data was
received or not, similarly to what's done already for TRXDv1 NOPE.ind in
TRXD and TCH channels in OsmoBTS. The aim at this change is to be able to
track TDMA clock in an accurate way without hops, and hence be able to
detect on time whether expected UL blocks (SF, RRBP poll) didn't arrive.
Older osmo-pcu versions can cope well with this change, they will simply
print an error upon ach data_len=0 messages received and submit a GSMTAP
block, then discard it, so tests still pass.
Nevertheless, a new module parameter is added to disable this new
behavior in order to avoid logs and pcap files ending up clogged with
uneeded information until a new osmo-pcu release appears.
Related: OS#5020
Change-Id: Ib4f97a9bcfa68230945effeb6412218faa64ec78
All BSS-originated BVC-Reset for PTP BVC must contain a Cell Identifier.
This test is not respecting that rule
libosmogb correctly detects this and discards the RESET, which is not
what is intended here:
20210305120722368 DLBSSGP ERROR BSSGP BVCI=1235 Rx RESET missing mandatory IE (gprs_bssgp.c:392)
Change-Id: I57ebfc50befc47d0690021d73b52c504f0df4f9a
Related: OS#3879
Fixes: Iea5519b453dd544c5a78b3b16ddb5132d20cf67c
This test currently fails to pass in master osmo-pcu (and latest) due to
T3169 not being implemented exactly as per specs (due to limitations in
detecting lost UL blocks with assigned USF).
Related: OS#5033
Change-Id: I56177850f084cdaf4fcac63ebdcdff9cef4e7a5d
They are tested together since anyway in order to reach T3191 we need to
go through X2301 (IDLE TBF timeout).
Related: OS#3928
Change-Id: Ib6dfc5711b9c6f1fd404bce424bbf4b115fc930e
Sometimes the DL data may arrive too late to PCU and it may be requested
to tx before it arrives, hence the PCU will in that case schedule and
transmit a UL ACK/NACK instead of the expected DL data.
Change-Id: Iaee546e2021e86ca6da19ab73cc8d283a827a665
All current test cases are for those 2 dialect. In preparation
of SNS test cases ensure those testcase have the correct dependency.
Change-Id: Ifda066553eabb0ace4aca12e60797c4950fcc498
As can be seen [1], it happens quite often that a test case gets
stuck and runs forever. Most of the existing test suites have
the guard timeout to prevent this. Let's make use of it here too.
[1] https://jenkins.osmocom.org/jenkins/view/TTCN3/job/ttcn3-remsim-test-latest/190/
Build has been executing for 1 day 23 hr!
Change-Id: I3897efd2a97c3e0d487057aa7bdc2967f3424dd8
Some modules from library depend on these native functions.
Without them the compilation fails during linkage:
Native_Functions.so: undefined reference to `Native__Functions::f__str__tolower(CHARSTRING const&)'
Native_Functions.so: undefined reference to `Native__Functions::f__inet__hntoa(OCTETSTRING const&)'
Native_Functions.so: undefined reference to `Native__Functions::f__inet__haddr(CHARSTRING const&)'
BSSMAP_Templates.so: undefined reference to `Native__Functions::f__inet6__addr(CHARSTRING const&)'
BSSMAP_Templates.so: undefined reference to `Native__Functions::f__inet__addr(CHARSTRING const&)'
Native_Functions.so: undefined reference to `Native__Functions::f__str__toupper(CHARSTRING const&)'
Native_Functions.so: undefined reference to `Native__Functions::f__inet__ntoa(OCTETSTRING const&)'
collect2: error: ld returned 1 exit status
make[1]: *** [BSCNAT_Tests] Error 1
Change-Id: I81227db58850375f0ce48e5c6e013381291a5904
Since we have new releases in -latest, we don't need this param since
it's never set to true.
Related: OS~5042
Change-Id: Ic496407fd2139b3a5221c30f1798797320cdee24
Harald Welte
Vadim Yanitskiy
Pau Espin
Alexander Couzens
Oliver Smith