|Age||Commit message (Collapse)||Author||Files||Lines|
Will be used in upcoming inter-BSC handover.
Do not invoke OSMO_STRINGIFY() with arbitrary names, just quote instead.
The idea was that OSMO_STRINGIFY() avoids typos by ensuring well-defined names
are stringified, but this highlights that OSMO_STRINGIFY() is in fact usable
with completely arbitrary arguments and actually lacks the validation part :/
Add lchan and timeslot FSM charts to illustrate planning of how osmo-bsc should
handle lchan assignment and release.
Modify assignment, handover, lchan-release charts according to the new plan.
Another small step towards being aware of the subscriber identity.
Any connection initiated by paging will subsequently log the subscriber's
identity -- of course not necessarily the IMSI, if paging was done by TMSI.
This is only for Paging, not the Paging Response; for that see, L3 Complete.
This is a tiny step towards being aware of a connection's subscriber identity.
Iff the Layer 3 Complete message contains an IMSI, associate a bsc_subscr with
the conn, so that subsequent logging and possibly meas_feed contains the IMSI.
For any L3 Complete using TMSI, this has no effect whatsoever.
Keep the bsc_subscr associated with the conn as long as possible, to benefit
logging, which then contains the subscriber identity (if any).
Instead of silently setting the use count to 1, instead increment with an
explicit bsc_subsct_get(), which then logs the event along with the place that
created the subscriber.
The guts of bssap_speech_from_lchan() and lchan_to_chosen_channel() have been
moved to libosmocore; call those instead.
The return value of bssap_speech_from_lchan() used to be -1 on error, now the
error value is 0. The only caller did not handle -1 properly, but fed it
directly to a uint8_t.
On gsm0808_chosen_channel() error, log the error. Proper handling is missing.
Fixing the error handling in send_ass_compl() is a separate issue: currently it
is limited to logging, there is no way to return an error yet, nor any actions
to take on error.
Depends: Icca23940791f97fa64dbc3f2734270b99f9550c1 (libosmocore)
In each code path within rsl_rx_chan_act_nack(), do separate logging of the
NACK to ensure proper termination of each log line.
When receiving a Chan Act NACK, we possibly mixed a LOGP() within an
unterminated other LOGP() that had not been ended with a LOGPC() yet.
When the BTS responds with a Chan Act NACK, i.e. the lchan could not be
activated, immediately signal Assignment Failure to the MSC (in
In handle_chan_nack(), adjust log: instead of waiting for timeout, we now
signal Assignment Failure.
Drop misleading logging from bsc_assign_fail(): instead of transmitting the
Assignment Failure message, it actually signals an FSM event. Leave logging of
that to the FSM logging.
Call one common function to signal Assignment Failure and transition back to
struct bsc_api was used to provide an abstract API for both osmo-bsc and
osmo-msc, between BSC and MSC. That's no longer needed, so get rid of it, to
prevent code turbulences it creates for no reason.
Move all of libbsc/ into osmo-bsc/, and separate/move some implementations to
allow linking from utils/* and ipaccess/* without pulling in unccessary
Some utilities use gsm_network and gsm_bts structs, which already include data
structures for fairly advanced uses. Move initialization that only osmo-bsc
needs into new bsc_network_init() and bsc_bts_alloc_register() functions, so
that the leaner tools can use the old gsm_* versions without the need to link
everything (e.g. handover and lchan alloc code).
In some instances, there need to be stubs if to cut off linking "just before
the RSL level" and prevent dependencies from creeping in.
- abis_rsl_rcvmsg(): the only program currently interpreting RSL messages is
osmo-bsc, the utils are merely concerned with OML, if at all.
- paging_flush_bts(): ip.access nanobts models call this when the RSL link is
dropped. Only osmo-bsc actually needs to do anything there.
- on_gsm_ts_init(): the mechanism to trigger timeslot initialization is related
to OML, while this action to take on init would pull in RSL dependencies.
utils/ and ipaccess/ each have a stubs.c file to implement these stubs. Tests
implement stubs inline where required.
From src/utils/, src/ipaccess/ and tests/*/, link in .o files from osmo-bsc/.
In order for this to work, the osmo-bsc subdir must be built before the other
source trees. (An alternative would be to include the .c files as sources, but
that would re-compile them in every source tree. Not a large burden really, but
unless linking .o files gives problems, let's have the quicker build.)
Minor obvious cleanups creep in with this patch, I will not bother to name them
individually now unless code review asks me to.
1) libbsc has been separate to use it for osmo-nitb and osmo-bsc in the old
openbsc.git. This is no longer required, and spreading over libbsc and osmo-bsc
2) Recently, ridiculous linking requirements have made adding new functions
cumbersome, because libbsc has started depending on osmo-bsc/*.c
implementations: on gscon FSM and bssap functions. For example, neither
bs11_config nor ipaccess-config nor bts_test need handover_cfg or BSSMAP
message composition. It makes no sense to link the entire osmo-bsc to it, nor
do we want to keep adding stubs to each linking realm.
The TEST_SCAN_TO_MSC is completely unused, but still the code linked
bsc_scan_bts_msg() for the unused code path. This is unlikely to ever be
expanded. Remove TEST_SCAN_TO_MSC and reduce linking requirements.
(I am this close to dropping the test entirely to avoid continuous linking
annoyance, but ok, since nothing else tests timezones AFAIK, keeping it.)
Rationale: channel_test is trivial and useless, probably only exists as a
legacy from openbsc.git.
- it tests two printf()s of "(bts=45,trx=0,ts=3,ss=4)", hardly useful.
- it tests ts_subslots() behavior for dyn ts, which will soon be replaced by a
most trivial mapping that does no longer need complex dyn ts state
evaluation (when introducing the new timeslot FSM to handle dyn TS switchover).
- All it does is compose four Paging Requests and feed it into the osmo-bsc
bssap_udt, and expects the cell identifier to be decoded properly.
- To do so it employs a comparatively huge linking effort. This linking effort
is becoming annoying in my ongoing work and is just not worth it.
- We have comprehensive cell identifier tests in ttcn3, verifying that the
proper cells are paged, and that actually for *all* cell identifier types.
Only count paging responses which can be mapped to an active paging.
Unsolicited paging responses which do not correspond to an active paging
would increment the 'paging response' counter. This means the number of
paging attempts could be smaller than the number of paging responses,
which can look confusing in the 'show statistics' VTY command.
If we receive a BSSMAP ASSIGNMENT REQ for a speech channel
containing only a CIC but no AoIP transport layer address, then that's
illegal and must be rejected.
In Change-Id If362a0084de452727cd063063dfb645eca2f9beb we re-introduced
accepting CIC-only ASSIGNMENT REQ, but we failed to verify that this
actually only happens over a SCCPlite A interface.
This fixes the BSC_Tests.TC_assignment_cic_only testcase.
The user might not want to enable LCLS support for administrative
reasons. So let's keep it disabled by default, until somebody
explicitly enables it with "lcls-mode mgw-loop".
In the future, we may want to introduce a "lcls-mode bts-loop" where we
don't loop at the BSC-colocated MGW, but where we instruct the BTSs to
feed the RTP directly to each other. This would require a
falt/transparent IP routing between the BTSs in the RAN.
This code contains the following code:
* receive/parse/interpret LCLS specific BSSMAP IEs and PDUs
* osmo_fsm handling the various states and their transitions
* call leg correlation (finding the other subscr_conn with same GCR)
* communication between the two call-leg LCLS FSMs
* detection of supported / unsupported LCLS configurations
* display of GCR / LCLS information in "show conns"
* switch the media streams locally using MDCX to the MGW
We now print information such as
* SCCP connection ID
* MSC number
* handover decision2 fail count
* channel mode (SIGN/SPEECH)
* MGW endpoint
* secondary lchan (if any, e.g. during assignment)
* don't crash if conn->lchan == NULL
Also change format of related debug message to match similar format than
In doc/, add two message sequence charts with lots of implementation specifics,
to clarify the current code state. Mark various problems in red notes.
This chart and others should help to illustrate future code changes as I go
along refactoring handover, lchan allocation and adding inter-BSC handover.
Right now, it's impossible to see any ACC Ramping information unless RSL
category is set to DEBUG. Barring and Allowing Access Control Class is
an important event which should be printed in most cases.
Increase log levels of messages printed during some error conditions to
be handled as errors.
It's not a good idea to keep extern declarations copied over
half a dozen C files. Let's move it to a header.
The bsc_msc_connection dates back to the old pre-libosmo-sigtran
days, and 90% of the field members weren't used at all (even the
new sigtran specific ones!). Let's merge what remains into struct
As a side effect, the already dysfunctional "dest A.B.C.D" VTY
command has been removed from the MSC node.
There's quite a bit of fall-out in the CTRL interface, which was
the code with strongest ties to bsc_msc_connection. This was
resolved by properly porting CTRL handling over to libosmo-sigtran,
meaning that an IPA/SCCPlite connected MSC can now again send CTRL
GET/SET commands, and can also receive those selective few TRAPs
that old osmo-bsc-sccplite also sent to its MSC[s].
osmo-bsc_nat is too heavily tied into legacy SCCPlite code, as it
is not using libosmo-sigtran/osmo_ss7 so far. It's also full of
customer-specific code and it's shared use of some libbsc code here
has been complicating osmo-bsc development.
The current plan is to continue to use osmo-bsc_nat from openbsc.git
for those legacy users that need it, and not use osmo-bsc_nat in
new 3GPP AoIP setups. Should we ever get a strong demand for an AoIP
based bsc_nat, we can still revisit this later.
In IPA/SCCPlite, the MSC is handling MGCP to the BSC-MGW directly
and the BSC only takes care of the BTS-side MGCP connection. We
achieve this by transitioning directly from WAIT_MDCX_BTS into ACTIVE
in this case.
In IPA/SCCPlite, we have to use the CIC to construct the MGCP
endpoint name instead of the usual dynamic endpoint allocation.
Also: Move mgcp_timeslot_to_port() next to it, as they are
more or less the inverse transformation of each other.
We used to have hard-coded M3UA. Let's allow the user to configure
this per MSC using a new "asp-protocol (m3ua|sua|ipa)" VTY command.
For SUA this should just work 1:1 without any trouble. For IPA,
this of course only changes the underlying transport without reflecting
the various differences in terms of BSSMAP ASSIGNMENT, MGCP handling,
In gscon_fsm_wait_mdcx_bts() we try to allocate conn->user_plane.fi_msc
but then check whether conn->user_plane.fi_bts is set, possibly due to
a copy+paste error. Let's fix that.
Use memcpy() to avoid unaligned access, instead of writing through a
pointer cast to uint32_t. Problem spotted by address sanitizer:
abis_nm.c:2802:24: runtime error: store to misaligned address 0x7ffc95396706
for type 'uint32_t', which requires 4 byte alignment
0x7ffc95396706: note: pointer points here
81 0b bb 80 00 00 00 00 ed 79 28 56 00 00 e0 9c 00 00 a0 61 00 00 ...
The state ST_WAIT_MODE_MODIFY_ACK can never be reached by the
current FSM model.
- Remove ST_WAIT_MODE_MODIFY_ACK and all related code
* Check the message length once at the start, before any other actions.
* Use only one local gsm48_hdr pointer.
* Read the cause value once near the top, re-use it.
* Log "ASSIGNMENT COMPLETE" always, not only during handover.
* Fully initialize local struct lchan_signal_data.