Age | Commit message (Collapse) | Author | Files | Lines |
|
There's no real need to use -1 to indicate echo timer as disabled, since
0 can also be used (it doesn't make sense to have a timer timeout of 0).
This way code is simplified.
Change-Id: I689034887188a53590eddeffda781629694eb5ed
|
|
When a MS MM state is READY its exact location is known (PCU).
On Gb, T3314 (aka TS 23.060 "READY timer") sets the MM state from
READY to STANDBY, where only the RA is known.
Introduce a second set of timer variables, because state timer
can run while another packet state timer is timing out.
Related: OS#1941
Change-Id: I4ce23ebe50d141076c20c9c56990b7103cd25e55
|
|
Recent commit added an assert to make sure unexpected conditions were
happening in sgsn_mm_ctx_cleanup_free(). Old code was passing
mm->gb.tlli to gprs_llgmm_assign with "new tlli" being all-1's (aka
unassign mm->gb.tlli).
The commit changed the code to use gprs_llgmm_unassign, which uses
llme->tlli instead of mm->gb.tlli, and the assert was used to make sure
no behavior change occured with the commit.
It seems TTCN3 test TC_attach_auth_id_timeout triggers that assert, and
after closer debug it seems mm->gb.tlli == llme->old_tlli, which makes
sense since there's a mm->gb.tlli_new which is expected to be
llme->tlli.
When TLLI changes in GMM (Attach Request or RA Update), it is stored
into mm->gb.tlli_new and assigned on the LLC layer using gprs_llgm_assign(),
and upon completion signalling from MS, (after handling response to initial request)
it is assigned to mm->gb.tlli (and value kept in mm->gb.tlli_new).
So mm->gb.tlli and mm->gb.tlli_new usually contain the same value unless
a new TLLI is allocated, and during the span of
Request->Response->Complete it is kept different, the LLC layer having assigned
the value of mm->gb.tlli_new.
So, old code (before the commit adding the assert) was wrongly using
mm->gb.tlli instead of mm->gb.tlli_new at the moment of unassigning (but
not really problematic in practice since behavior is the same as long as
"old TLLI" value is not all-1's.
So we are fine and correct using gprs_llgm_unassign() (which passes llme->tlli
as "old TLLI") instead of what used to be done before.
In any case, the expected behavior is to free the llme object and get
rid of everything...
Fixes: 788863cda53298c24110d0fe0f8cd3309cdec747
Change-Id: I482acdbdf05ce0cb0a5804206672512854067f5b
|
|
May be useful to detect unexpected conditions which could end up in
memory leaks.
Related: OS#3957
Change-Id: I0d175501083ce458ff1c07ad38761d2cbf4ea470
|
|
Change-Id: Ib8be5af2a5e92a7403505b73ce4c1751832de40d
|
|
GCC 9 complains that variable 'gsm_cause' in do_act_pdp_req() may
be uninitialized. This may happen if sgsn_mm_ctx_find_ggsn_ctx()
would return NULL due to no static GGSN configured.
Change-Id: I09c608045dd35b9898b82e236a306ab9a6c2c0b9
|
|
Change-Id: Id1511c5022a239db5d0b44ec7adf048cca307751
|
|
Depends: libosmocore I52b9f6b5f3e96d85a390ba2af21d7814df8aaeec
Change-Id: Icf9f466efce520779c926b47b6e6d6c9815120eb
|
|
Previous commit introduced command "authentication (optional|required)",
which is only meaningful if auth-policy is remote. Upon adding the cmd,
it changed the default logic for remote policy to not require
authentication, which broke TTCN3 tests because sgsn no longer tries to
authenticate the users.
Since it's actually good to enable authentication by default where
possible, let's enable it by default when on auth-policy remote.
In order to do so, let's simply not care about the value of variable
require_authentication if auth_policy is not REMOTE. As a result, we
drop parts of the previous patch and remove unneeded checks (which are
only partially useful based on order of commands during VTY read).
Fixes: 794f446a284ed1ac6d31eb79a8f4c874d66fc34e
Change-Id: Ic707a95af178b44f08809df3d3bc8354bf34273c
|
|
Commit 176a4d2f33865a5c0433f8679f5e68f209d7b874 moved echo timer related
code to its own function but did some mistakes when moving the logic
from several places into its own function. As a result, echo timer was
only enabled after the 2nd pdp ctx was created, instead of the expected
1st.
First, let's be consistent and always call the function *after* changing
state, since that's what the function expects. This fixes the issue.
Finally make the logic in the function more intuitive by checking in the
if clause the only case where actually the echo timer should be enabled:
Only if policy specifies so and we have at least 1 pdp ctx against that ggsn.
Fixes: 176a4d2f33865a5c0433f8679f5e68f209d7b874
Change-Id: I826030978edb61ea5a172c2b72f63758206a6246
|
|
Change-Id: I4a83c5799f0dbd5eb762039c6cfba671f6e465be
|
|
Change-Id: I93f0dc721c2eff8a87fb9248882f24768f708713
|
|
Move the check of the echo timer into an own function.
The gtp echo timer must be re-check everytime the
echo-timer has been modified or deactivated via vty.
Fixes the TTCN3 SGSN_Tests.TC_attach_restart_ctr_echo
Change-Id: Ia33471a9a9cfc3887facb665c82094b99932052a
|
|
When the GGSN crashs, the SGSN will be notified after
it comes back. Because of the async operation,
the mm ctx could be already gone.
Change-Id: I507a8c2193c84f8dff7f5d669adcd3583331f289
|
|
The old GMM Attach Request handling used a recursive function
which can not handle certain states and is quite complex and hard to
extend.
The new FSM handles such request in a FSM and can be called multiple
times.
Change-Id: I58b9c17be9776a03bb2a5b21e99135cfefc8c912
|
|
Change-Id: Ic638849c6687c376c4c0c36cc286d499a073d6ca
|
|
When PDP CTX CREATE ACK is received with an increased RestartCtr, cb_recovery2
is called first, which will dettach ggsn from al pdp ctx (free the
pdp_t). But when giving control back from the ctrl, libgtp still uses
that freed ctx and sends it back to osmo-sgsn through cb_conf().
As specs state in any case that we need to handle the message containing
the increased RestartCtr as valid, we then need to avoid freeing the pdp
ctx and leave handling for later in cb_conf.
Depends: osmo-ggsn (libgtp) Change-Id I53e92298f2f6b84d662a3300d922e8c2ccb178bc.
Change-Id: I0989c00e18ca95a099e1a312940eaac71957b444
|
|
This commit fixes TTCN3 sgsn test TC_attach_pdp_act_user_deact_mt.
Change-Id: I204209c017aac8a8402cbb8d0a0200540abcc954
|
|
Scenario and behaviour before this commit:
- Received Echo Reply from GGSN has incremented RestartCounter
- func sgsn_ggsn_ctx_drop_all_pdp() is called to dettach all pdp ctx
from GGSN and request the MS to deact all related ctx.
- DEACT ACCEPT is received from MS, and then it tries to send DEL PDP CTX
to GGSN, expecting to receive a Confirmation and only then freeing the
pdp ctx.
The problem is that since the initial cause of triggering was a GGSN
restart, the GGSN doesn't know anything about that pdp ctx anymore, so
it's not useful sending it. We can instead dettach the GGSN and libgtp
ref at drop_all_pdp() time and then when we receive DEACT ACCEPT from MS
we can free the pdp ctx directly.
Change-Id: I1c74098e181552c218e152bf4ac5035cea770428
|
|
According to 3GPP TS 24.008 Section 6.1.3.4, the tear down indicator IE
maybe included in the DEACTIVATE PDP CONTEXT REQUEST message in order
to indicate whether only the PDP context associated with this specific
TI or all active PDP contexts sharing the same PDP address and APN as
the PDP context associated with this specific TI shall be deactivated.
As we don't permit/support establishing multiple PDP contexts using
the same APN and PDP address, it shouldn't really make any difference.
Nevertheless, we want to clear everything, so let's include it.
Change-Id: Ia9bc2d0e93362a8473eac5cf4c7e8ffa41c79e5b
|
|
In sgsn_pdp_ctx_terminate, a pdp ctx is terminated and the mm ctx is
detached. However, T3395 may still be armed and then pdpctx_timer_cb
will trigger, and attempt to use the pdp->mm ctx which was already
detached (set to NULL) when calling
gsm48_tx_gsm_deact_pdp_req()->mmctx2msgid().
Following list of log lines shows the scenario+crash, in which osmo-sgsn
is trying to deactivate the ctx all the time but the PCU doesn't ACK it,
and then at some point the PDP context is forced released.
osmo-sgsn/src/gprs/gprs_gmm.c:2294 MM(901700000015254/d7e9ab95) <- DEACTIVATE PDP CONTEXT REQ
osmo-sgsn/src/gprs/gprs_gmm.c:1464 MM(901700000015254/d7e9ab95) -> GMM DETACH REQUEST TLLI=0xd7e9ab95 type=GPRS detach Power-off
osmo-sgsn/src/gprs/gprs_gmm.c:313 MM(901700000015254/d7e9ab95) Cleaning MM context due to GPRS DETACH REQUEST
osmo-sgsn/src/gprs/gprs_sgsn.c:332 MM(901700000015254/d7e9ab95) Dropping PDP context for NSAPI=5
osmo-sgsn/src/gprs/gprs_sgsn.c:434 PDP(901700000015254/0) Forcing release of PDP context
osmo-sgsn/src/gprs/gprs_sndcp.c:508 SNSM-DEACTIVATE.ind (lle=0x62100001bca0, TLLI=d7e9ab95, SAPI=3, NSAPI=5)
osmo-sgsn/src/gprs/sgsn_libgtp.c:310 PDP(---/0) Delete PDP Context
osmo-sgsn/src/gprs/gprs_gmm.c:2294 MM(---/ffffffff) <- DEACTIVATE PDP CONTEXT REQ
osmo-sgsn/src/gprs/gprs_gmm.c:305:25: runtime error: member access within null pointer of type 'const struct sgsn_mm_ctx'
Program received signal SIGSEGV, Segmentation fault.
0x0000555555698c1b in mmctx2msgid (msg=0x61d0000172e0, mm=0x0)
at /home/pespin/dev/sysmocom/git/osmo-sgsn/src/gprs/gprs_gmm.c:305
305 msgb_tlli(msg) = mm->gb.tlli;
(gdb) bt
#0 0x0000555555698c1b in mmctx2msgid (msg=0x61d0000172e0, mm=0x0)
at osmo-sgsn/src/gprs/gprs_gmm.c:305
#1 0x00005555556b170a in _gsm48_tx_gsm_deact_pdp_req (mm=0x0, tid=0 '\000',
sm_cause=38 '&')
at osmo-sgsn/src/gprs/gprs_gmm.c:2297
#2 0x00005555556b1a2e in gsm48_tx_gsm_deact_pdp_req (pdp=0x6140000008a0,
sm_cause=38 '&')
at osmo-sgsn/src/gprs/gprs_gmm.c:2311
#3 0x00005555556b876c in pdpctx_timer_cb (_pdp=0x6140000008a0)
at osmo-sgsn/src/gprs/gprs_gmm.c:2717
#4 0x00007ffff355eb3e in osmo_timers_update ()
at libosmocore/src/timer.c:257
#5 0x00007ffff356255c in osmo_select_main (polling=0)
at libosmocore/src/select.c:254
#6 0x00005555556f17cb in main (argc=3, argv=0x7fffffffe298)
at osmo-sgsn/src/gprs/sgsn_main.c:531
Change-Id: I2120e53ade6cabad37f9bd99e6680a453411821b
|
|
Before this commit, echo req/rsp logic was implemented in libgtp but
never used in osmo-sgsn.
This commit adds a timer which periodically sends a GTP ECHO Request to
every GGSN if there's at least one pdpd context associated with it. This
way by checking the restart counter in the ECHO Reply it can be known if
the GGSN was restarted. In this case, logic already present in osmo-sgsn
will terminate all pdp contexts associated with that GGSN.
Change-Id: I9d714726785407859f26bbef052cd0efc28e8dae
|
|
This way we can easily track all pdp context associated to a specific
ggsn, which is useful to handle some scenarios, such as the one
implemented in next commit, in which specs references that GSNs should
ping only other GSNs with at least one pdp ctx in common. So the list
of pdp ctx per GGSN is really useful too (and cheap computationally)
to check if we should arm or disarm the echo procedure timer.
So this commit can be seen as a preparation for next commit.
Change-Id: I3bbcc0883df2bf1290ba8d4bd70db8baa494087a
|
|
This avoids potential licensing incompatibility and makes integration of
Debian packaging patches easier.
The libosmocore version requirements are fine already but for jenkins
tests to pass we have to have Ic77866ce65acf524b768882c751a4f9c0635740b
merged into libosmocore master.
Related: OS#1694
Change-Id: I2b687b7f07ef05bbd861b8479cad5a958a3dde92
|
|
Due to recent libosmocore's change we can't allocate rate counters with
the same name and index which are already allocated. This causes
sgsn_mm_ctx_alloc_iu() failure for multiple subscribers.
Fix this by using conn_id parameter from ranap_ue_conn_ctx.
Change-Id: I1062ffdcac96c82269cab6f4e7ae50e28dc3aa44
Related: OS#2757
|
|
'.' is an illegal character in counter names, as they are exported
via CTRL interface, where '.' has a special meaning that cannot be
used by strings comprising the variable name.
Change-Id: Ie7734cc42151581897d220b445984448ceb57aed
|
|
'.' is an illegal character in counter names, as they are exported
via CTRL interface, where '.' has a special meaning that cannot be
used by strings comprising the variable name.
Change-Id: Iec382ec4ee54beb2937431f5a9d8d1171224eebb
|
|
Change-Id: I281ef585fffc2644682c8282224fb1c2da5ca795
|
|
Change-Id: I5d27ff93e56cd13e0e70edd15e2080201e35e91f
|
|
Remove libiu here, use the functions from libosmo-ranap instead, by applying
the ranap_ / RANAP_ prefix.
Corresponding change-id in osmo-iuh.git is I6a3f7ad15be03fb94689b4af6ccfa828c25f45c0
To be able to run the msc_vlr tests for RAN_UTRAN_IU without Iu client headers
available, add iu_dummy.h, containing mere function signatures that match
iu_dummy.c and a mostly empty struct ranap_ue_conn_ctx.
Make sure we can build with and without --enable-iu: include osmo-iuh headers
only with --enable-iu.
Change-Id: Ib8c4fcdb4766c5e575618b95ce16dce51063206b
|
|
In 2015, Jacob moved/copied related functions to libosmocore, but
for some reason didn't remove the copies here. Let's follow-up on
that and remove duplicated code.
The libosmocore commit introducing osmo_apn_to_str() was
8114294bf29ac6e44822c0ae43d4b0819f11b022
Change-Id: I7315ffcbed8a54cca2056f313bb7783ad82d0ee9
|
|
In case the counter group allocation fails, we must handle this
gracefully and fail the allocation of the parent object, too.
RelateD: OS#2361
Change-Id: I7dad4a4d52fe05f6b990359841b4408df5990e21
|
|
For some GGSNs we need to insert the PDP Charging Characteristics
that were returned. We receive these values from GSUP and will
fill them into the tlv structure when finding the ggsn context.
Change-Id: I1725bfd2403d29ce3550bfcd6fcc1498426ef906
|
|
osmo-nitb becomes osmo-msc
add DIUCS debug log constant
add iucs.[hc]
add msc vty, remove nitb vty
add libiudummy, to avoid linking Iu deps in tests
Use new msc_tx_dtap() instead of gsm0808_submit_dtap()
libmgcp: add mgcpgw client API
bridge calls via mgcpgw
Enable MSC specific CTRL commands, bsc_base_ctrl_cmds_install() still needs to
be split up.
Change-Id: I5b5b6a9678b458affa86800afb1ec726e66eed88
|
|
This is the first step in creating this repository from the legacy openbsc.git.
Like all other Osmocom repositories, keep the autoconf and automake files in
the repository root. openbsc.git has been the sole exception, which ends now.
Change-Id: I9c6f2a448d9cb1cc088cf1cf6918b69d7e69b4e7
|