A CoDel state per prio queue is needed, otherwise the sojourn time state
is not properly reset when a high prio packet is dequeued.
If we have a global codel state shared for all prio queues of an MS, then
basically high prio (GMM) packets will "never" be dropped because they are
handled/dequeued way quicker, so it's sojourn time will be below the
threshold most probably, stopping the "dropping" state for the rest of
lower prio packets.
The handling of different codel states is moved from MS object to the
llc_queue, also offloading already loaded dl_tbf.cpp in the process.
This will also allow in the future setting different CoDel parameters
for different priority queues if needed.
Tests need to be adapted since now the CoDel and PDU lifetime are
incorporated into the llc_queue_dequeue(). Also because dequeue now also
accesses related MS fields.
Related: OS#5508
Change-Id: I2bce2e82ab6389d8a70130a5c26a966a316b0fa4
There's no real reason (other than historical) why code in llc should be
kept as c++ code. Let's rewrite it as C so that it can be included by
existing C code without having to add C->C++ shim, ifdefs all around,
etc.
This simplifies code and easies modification/improvement of the related
objects.
Change-Id: I250680ba581167d7398b2f734769c756cbb61c48
Remove the paragraph about writing to the Free Software Foundation's
mailing address. The FSF has changed addresses in the past, and may do
so again. In 2021 this is not useful, let's rather have a bit less
boilerplate at the start of source files.
Change-Id: I4a49dbeeec89b22624c968152118aecf8886dac6
We basically want to probe whether it's possible to allocate TBFs, or
whether we know it will fail due to all main resources being already in
use (TFI, USF).
Having bts_all_pdch_allocated() return false doesn't mean though that an
MS will be able to allocate a TBF for sure. That's because further
restrictions are applied based on MS: whether it was already attached to
a specific TRX, whether the ms_class allows for a certain multislot
combination, etc. However, it should provide a general idea on whether
for sure the PCU is unable to provide more allocations. More fine
grained state about failures can still be followed by looking at
tbf:alloc:failed:* rate counters.
Related: SYS#4878
Depends: Iabb17a08e6e1a86f168cdb008fba05ecd4776bdd (libosmocore)
Change-Id: Ie0f0c451558817bddc3fe1a0f0df531f14c9f1d3
The whole paging path and data structre is cleaned up.
New MS helpers ms_imsi_is_valid() and ms_paging_group() are introduced
to help in the process and keep implementation details inside GprsMs
class.
Related: OS#5303
Change-Id: I4c0838b26ede58e4b711410eee2a8e4f71e9414b
Log before the action is done ("Detaching", not "Detached"). This way,
if something crashes we see that last line.
Change-Id: I6811d3772e2ac850741d3db5ec32be5c0812e81a
This commit actually addresses 2 errors:
1- gprs_bssgp_pcu_rx_paging_ps() called gprs_rlcmac_paging_request()
with MI which can be either TMSI or IMSI, and the later always called
bts_pch_timer_start() passing mi->imsi regardless of the MI type. Hence,
trash was being accessed & stored into bts_pch_timer structures if MI
type used for paging was TMSI.
2- When the MS received the PS paging on CCCH and requests an UL TBF, it
will send some data. If one phase access is used for whatever reason,
the IMSI may not be yet available in the GprsMs object since we never
received it (and we'd only have it by means of PktResourceReq). Hence,
let's better first try to match the paging by TLLI/TMSI if set in both
places, and otherwise use the IMSI.
Related: OS#5297
Change-Id: Iedffb7c6978a3faf0fc26ce2181dde9791a8b6f4
This allows distinguishing when a TBF didn't set the TFI. Useful to
identify dummy reject TBFs, etc, and make sure a non-dummy TBF set its
TFI properly.
Change-Id: Iecf54a24041bd14f4ef5b86e57c3732e1b69d463
Let's set them to match expectancies for this type of dummy TBFs, in
order to avoid acidental use/access of other timeslots to the one where
the reject was associated to.
Also use tbf_assign_control_ts() to log the TS used for the TBF, similar
to what's used in other places where control_ts is assigned.
Related: OS#5293
Change-Id: I32dcb29ad24519082b8665921efcce0b5a16d12e
These new counters allow the user to find out which kind of access are
MS requesting.
Related: SYS#4878
Change-Id: Id87c3a53d3acee92499987c843130e358f54742c
If we decode Exist bit as "1" but we are at the end of the message, and
all the Next items we'd read are expected to be possibly NULL, then swap
the Exist bit in the decoded structure as "0" in order to tell the
decoder user that the related information structure is actually unset,
as if "0" was received.
Related: SYS#5552
Related: OS#4955
Related: OS#5020
Change-Id: I38602e4b680ed87297c7e440691a494c07cad446
All additional release fields are considered optional, and the
CSN_DESCR for Content_t already marks almost all as such, except
DownlinkDualCarrierCapability_r7.
It has been found that some MS transmits a MS RA Capability with a Length=61 bits
where the last bit in the buffer is setting the Exist bit for
DownlinkDualCarrierCapability_r7 as 1. Hence, the CSN1 decoder failed to
decode the whole message because it expected to keep reading there
despite there's no more bytes to read.
While this is could actually be considered an MS bug, let's relax our
expectancies and simply consider the case { 1 <end> } as it was { 0 },
and mark skip decoding DownlinkDualCarrierCapability_r7. That waht
wireshark (packet-gsm_a_gsm.c) or pycrate do for instance.
This patch itself doesn't fix the problem where actually the Exist bit
is stored as 1 in the output decoded structure, but simply allows keep
ongoing with decoding until the end. This issue will be fixed in a
follow-up patch.
Related: SYS#5552
Related: OS#4955
Related: OS#5020
Change-Id: I9a2541bd3544802a646890f32725201836abb0da
This RA Cap creaes a decoding error on our CSN1 decoder, but seems to be
handled properly by wireshark's own decoder as well as pycrate.
The ending bit of last byte in "MS RA capability 1" has a "1" which
according to spec should flag the existance of
DownlinkDualCarrierCapability_r7, but nothing else comes after it. This
matches the expectancies as per Length field of the first RA Cap.
Related: SYS#5552
Related: OS#4955
Related: OS#5020
Change-Id: I51235e8575f4b992b44078713ec67bbccfd13293
This helps distinguishing the case where a TBF is in the initial state
and the unexpected case where osmo_fsm_inst_state_name reports "NULL"
due to fi pointer being NULL.
Change-Id: Ieaabfc9fa0dedb299bcf4541783cf80e366a88c3
This one is larger than some of the other already available.
The decoder is wroking as expected here.
Change-Id: I5d986f68395326f894349446194090b1ddaecd69
The test uses get_poll_fn() to submit a UL ctrl block on the next
expected poll TS+FN.
During initial transmit_dl_data(), 2 POLLs for DL_ACK are set on
different TS, but the test only updates the clock for one of them in
send_ul_mac_block(). As a result, one POLL item is left in the rb_tree,
and later on, when send_control_ack() is called, get_poll_fn() will pick
that 2nd DL_ACK poll FN+TS which was left untouched (due to sending no
events to the PCU clock) instead of the FN+TS which was expected to be
allocated by PCU for DL_ASS.
Until now this was not an issue since rcv_control_ack() was not properly
veirfying what the poll scheduler was expecting and accepted it. This is
no longer the case after the follow up patch refactoring
rcv_control_ack() and improving its robustness.
Change-Id: I3a4b089fe66a99e73e07bd1c690cd4d67752fad9
There's no real use in having those 2 methods separately, and only adds
complexity. Let's merge it to have 1 TBF code path handling DL ACK/NACK.
Change-Id: I546d2e46bda96a2f551b28673464e57831c71828
There are 2 methods "rcvd_dl_ack()" in osmo-pcu code. One is used by
osmo-pcu itself, and the other is only used in tests.
Changing the tests to use the same method as osmo-pcu allows removing
the second one, and with it, a lot of code and complexity out of
osmo-pcu.
Change-Id: I14d9312cb61534dc97fca83141b9c0cd933c9206
The side effect is that the timer is enabled for other scenarios where a
PACCH assignment happens, like an Assignment Reject or Ul Assignment
(that's why there's more lines showing up now in TbfTest.err).
Change-Id: Ib8ab2f7397ad05c6fcd5dd74af55a1e2c56e1463
The flag is only used to print some non interesting stuff, let's drop it
in order to simplify code. We can add later whatever we want in the new
shiny FSM.
Change-Id: I13f92f058c219f230d57b3c00b8ae1d187603813
The flag is only used to print some uninteresting stuff, let's drop it
in order to simplify code. We can add later whatever we want in the new
shiny FSM.
Change-Id: I20aa7f83cc4f32de129e64c74a91745b983a7b16
We are freeing the object immediately afterwards anyway, so no need to
pretend it went through the normal state release.
Leaving current state as it is actually provides more information on
what was the status/state at the time the TBF had to be freed.
Change-Id: I3016caaccc2c43e1e300f3c6042d69f8adcd9d69
PdchUlcTest output changes because the original state NULL is not
expected when transactioning to RELEASING upon MAX N310* being hit. In
any case, none of those events should happen in NULL state, but we
don't really care about TBF states there so we are fine with whatever
the state is.
Related: OS#2709
Change-Id: I516b8d989a0d705e5664f8aeaf7d108e0105aa16
While at it, method maybe_start_new_window is renamed to
rcvd_dl_final_ack to make more sense out of the code.
Related: OS#2709
Change-Id: Iebd650c1036ef2d5132789778be7117ce3391c01
Prior code was wrong, as in it did stuff diferent to what is expected
and actually done in osmo-pcu. In osmo-pcu code, the function is guarded
and only called in FLOW or FINISHED state by the scheduler.
Change-Id: If8029bee90adceee128ebb20c033756efd50e90e
At some point later in time the state_flags will most probably be split
into different variables, one ending up in a different FSM. It is moved
so far to the exsiting FSM from the C++ class since it's easier to
access it from C and C++ code, and anyway that kind of information
belongs to the FSM.
Related: OS#2709
Change-Id: I3c62e9e83965cb28065338733f182863e54d7474
Run bts_pch_timer_remove() on each entry of the BTS specific pch_timer
list, so we don't have a memory leak and so the timer doesn't
potentially fire for a deallocated BTS.
Fixes: d3c7591 ("Add counters: pcu.bts.N.pch.requests.timeout")
Change-Id: Ia5e33d1894408e93a51c452002ef2f5758808269
This reverts commit 846fd248dc.
The commit introduced a leak of UL-TBF, which do not time out and
accumulate indefinitely, leading to out-of-memory for the running
osmo-pcu process.
A proper fix for the leak is pending on a development branch pespin/fsm,
but that branch is not yet ready for merging. Hence let's re-introduce
timer T3169 to avoid the OOM due to lingering UL-TBF.
Related: OS#5209
Change-Id: I99a7d2ddf68a76739ce2db1d6a44967dd97667b0
When slowly debugging test_tbf_dl_llc_loss, bssgp_tx_llc_discarded() may
trigger, submitting events to the libosmogb code. Since it didn't
properly set up the callback, it would end up in a null pointer
dereference when lib code tried to use backward-compatible API (which
was neither set up properly).
"""
TBF(TFI=0 TLLI=0xc0123456 DIR=DL STATE=ASSIGN) Discarding LLC PDU because lifetime limit reached, count=3 new_queue_size=0
BSSGP (BVCI=2234) Tx LLC-DISCARDED TLLI=0xc0123456, FRAMES=3, OCTETS=57
/git/libosmocore/src/gb/gprs_ns.c:271:2: runtime error: member access within null pointer of type 'struct gprs_ns_inst'
"""
"""
(gdb) bt
#0 0x00007ffff729cac0 in gprs_active_nsvc_by_nsei (nsi=nsi@entry=0x0, nsei=2234, bvci=bvci@entry=0)
at /git/libosmocore/src/gb/gprs_ns.c:271
#1 0x00007ffff72b1fec in gprs_ns_sendmsg (nsi=0x0, msg=0x621000000160) at /git/libosmocore/src/gb/gprs_ns.c:1087
#2 0x00007ffff72d1803 in _gprs_ns_sendmsg (ctx=<optimized out>, msg=<optimized out>) at /git/libosmocore/src/gb/gprs_bssgp.c:80
#3 0x00007ffff730226f in bssgp_tx_llc_discarded (bctx=<optimized out>, tlli=<optimized out>, num_frames=<optimized out>, num_octets=<optimized out>)
at /git/libosmocore/src/gb/gprs_bssgp_bss.c:249
#4 0x000055555588243e in gprs_rlcmac_dl_tbf::llc_dequeue (this=0x7ffff1622860, bctx=<optimized out>)
at /git/osmo-pcu/src/tbf_dl.cpp:413
"""
Change-Id: Iee5bcf21afc8980a14f90f5b1ead6d2460a244ea
In existing previous code, pdch->ulc would be freed in
gprs_rlcmac_pdch::free_resources() when it became disabled as per PCUIF
info_ind (for instance, when a DYN TS is switched PDCH->SDCCH8).
However, pdch->ulc was so far only allocated during pdch_init, which is
only called during bts_alloc() time.
Hence, after first info_ind disabling it, if it became again enabled
(again by info_ind re-enabling it after SDCCH8 was not longer in use),
the pdch->ulc would be used again but it would point to freed memory.
Let's rearrange how/when resources are freed to make it more logical.
With this patch, pdch internal resources are freed upon ->disable(), and
re-allocated upon ->enable().
Change-Id: Id51f5f6a54ac9f24b784c17bc360ac38f5726fc7
This is only an initial implementation, where all state changes are
still done outside the FSM itself.
The idea is to do the move in several commits so that they can be
digested better in logical steps and avoid major break up.
Related: OS#2709
Change-Id: I6bb4baea2dee191ba5bbcbec2ea9dcf681aa1237
This is right now not an issue, but it will be whenever talloc
destructor contains extra steps like freeing an FSM.
Change-Id: I096ff56321c8ae5e66634537aae8b95804282c65
This allows more easily finding when this specific scenario happens, and
can easily be compared against the PACCH one.
Change-Id: I609792a40fda2a798ca71a0e9f5639d0a0f011d7
Store direction check to simplify the code.
Get rid of 2-step LOGP to avoid multi-row logs in gsmtap log.
Change-Id: Ia2e061da82ddce564b2d768d8ade1672c22934e2
Tbftest expectatins need to change because 5000/20 = 250 < 255, hence
the message is now sent as units of 20ms instead of seconds.
Related: OS#3928
Change-Id: I48b34b94b1a5dfb046a3a6cf8a0d944a7c9b6754
Document the function, make it look similar to usual TBF creation path
tbf_alloc_ul()->tbf_alloc_ul_tbf->tbf::setup(), which it mimics with
some differences.
Get rid of unneeded stuff like creating MS and settings its TLLI (that's
already done in only caller of the function). There's no need for
calling update_ms() either.
Change-Id: I61df2e4f0f0df1f8db941741a2d35a2319252c5e
Now that we finally handle N3101 and N3103 correctly, we can fix abuse
of T3169 we were doing to make sure TBFs were freed.
According to 3GPP TS 44.060, T3169 should be armed:
* N3101_MAX reached
* N3103_MAX reached
Furthermore, when T3169 is enabled, the tbf should be in state
RELEASING so that its USF is not used.
See full description: https://osmocom.org/issues/5033#note-2
Related: OS#5033
Change-Id: I2cec531e2633281b88f69ba065c0105580c81076
Use recently added PDCH UL Controller to verify expectancies.
Test test_packet_access_rej_prr is rewritten since it didn't make sense
as it was before, since it relied on osmo-pcu not checking stuff
properly to trigger the reject. The RACH requests are changed to
allocate 8 SBAs (maximum of 7 concurrent USFs). Allocating the SBA
doesn't reserve a USF, that happens at PKT RESOURCE REQUEST, hence we
end up exhausting resources there and triggering the REJECT at that
point.
Previous version of the patch allocated TBFs directly through RACH req,
and then submitted an extra PKT RESOURCE REQUEST which PCU didn't expect
to trigger the reject.
Change-Id: I157e72160317340ee7742c78c62a25d3d98fc01e
That field is not needed anymore, and it works only under the assumption
that only 1 poll request can be active at a time per TBF, which is not
true.
Change-Id: I9b8bed7741d385bab4cd8c64b841a78a02a05fe1
That field is not needed anymore, and it works only under the assumption
that only 1 poll request can be active at a time per TBF, which is not
true.
Change-Id: I63a34a702f028b871530fb7caeb13e8ea1cc78ac
Value 'false' is always passed by all callers of the function, so
there's no need to pass it. Furthermore, since it's false, there's no
need to access poll_fn since RRBP will always be invalid.
Change-Id: Ia48ce2a021865e76e813dedb22aca9c2522c5693
The poll_state logic was part of previous implementation (prior to pdch
ul controller) where the ssumption was that TBF could only had 1 POLL
request in transit, which is really not true. With current
infrastructure we don't need this state tracking at all.
Change-Id: Ie5b807ccd38aa736ae11b3310ca61ad0156ca4d4
There's no good reason to allow only for 1 concurrent POLL requested to
a TBF, it was onyl done this was as an implementation limitation factor.
It can well happen that several multiple POLLs may be in transit at the
same time, eg to get DL ACK/NACK as well as to get a CTRL ACK for a Pkt
Cell Change Continue (NACC).
Change-Id: Ic4080db684a4626cae90dd574d123081981284ca
This API is not really needed anymore, since anyway it works under the
assumption there can only be 1 POLL in transit per TBF, which isn't
necessarily true.
Change-Id: I875f51cade95faeb2d79dcebfead4c83e23a731b
This allows easily checking the initial reason to trigger the poll when
either it is received or times out.
Later on this reason can be transformed into an FSM event and sent to
the related FSM.
Related: OS#5020
Change-Id: Ie8fefd1f47ad674ce597a8065b15284088956bde
Make sure an unreserved FN is picked and reserved when allocating and
scheduling an SBA.
In practice this has no change in behavior right now, since anyway using
an offset of 52 FNs ensure no USF or POLL has alredy been scheduled that
far in the future. Since it's also impossible to allocate more than 1
SBA per PDCH and RTS FN, we are also safe about multiple SBAs being
allocated, because we use a hardcoded offset of 52.
However, that could change in the future, when we dynamically tweak the
current offset of 52 FN based on information from BTS about its AGCH
queue load:
* If load is high, we may need to increase the offset since it
will take more time for the BTS to transmit the TBF and hence we must
reserve a TBF starting time further in the future (higher FN).
* If load turns low, we may schedule next SBA a bit more nearby in time
than the previously allocated SBA, hence here there could be a
collision.
Related: OS#5020
Change-Id: I2d4e21e2307de6c17748e8da5c7e149c947a7eb9
This way PCU can now detect whether scheduled UL blocks through USF
were never received. This allows in a follow-up patch to start
increasing N3101 properly.
Related: OS#5033
Change-Id: Ia99c9edad6e5bd837e9baeb4fb2683b227887957
There's no need for setting the FN in RA.ind since we anyway already
receive a DATA.ind beforehand.
Furthermore, the applied delay of 5 in the call is not really used at
all.
Change-Id: I437f4f95d054aea96bec3b9343e495451020ff3c
Issue will be fixed in next commit. Leaving ASSERTs disabled so that
test passes in jenkins.
Related: OS#5020
Change-Id: I657db6b300363f8f3a9e4cfaf7a7f49e361a0512
TbfTest is updated to submit empty blocks to have somehow meaningful
output (at least as meaningful test results as before, not much). That's
because we must update bts->curr_fn to have polls expire.
Related: OS#5020
Change-Id: I683ca738ce5a133c49c36a1d94439a942d64a831
Right now we handle different types of UL allocations in different
classes like PollAllocator and SBAllocator, and they usually don't take
into account the other one in most cases. Furthermore, those objects are
usually per-BTS object, instead of per PDCH object.
This is a first step towards having a unified per-PDCH controller which
takes care of controlling what is scheduled and hence expected on the
uplink. Each PDCH has a UL Controller which keeps track of all reserved
uplink frame, be it SB, RRBP poll or USF assigned, all under the same
API.
As a first step, only the SBA part is fully implemented and used (being
it the easiest part to replace); TBF poll+usf will come in follow-up
patches later on. As a result, the SBAllocator per-BTS class dissappears
but some of its code is refactored/reused to provide more features to the
gprs_rlcmac_sba object, which is also further integrated into the new UL
Controller.
Related: OS#5020
Change-Id: I84b24beea4a1aa2c1528f41435f77bd16df2b947
Since recently (see Depends below), BTS side submits DATA.ind with len=0
to announce nothing was received on that UL block FN. This will allow
osmo-pcu track time more accurately, and use this information to quickly
find out if a UL block was expected as requested by RRBP or USF poll and
increment counters such as N3101 (finally being able to properly
implement timers such as T3619).
Depends: osmo-bts.git Change-Id I343c7a721dab72411edbca816c8864926bc329fb
Related: OS#5020
Change-Id: Ibc495173119465e74f726ddc36e312334e6dc0fd
If Idle TBF timer (X2031) is set to 0, it means the TBF release is
immediately started once all queued data has been scheduled. In that
case, we must set FBI=1 (by setting cv=0) and move to FINISH state.
This used to work over the usual path where X2031 != 0, because release
start will alays happen at a later sched poll time where a dummy LLC
frame is sent and FBI set accordingly.
Change-Id: Ib20602936ae084c413f6bfe14eea33b602020be0
Old commit getting rid of LLC UI dummy and updating create_new_bsn()
function introduced a bug by not moving update of value m_last_dl_drained_fn
prior to a new break introduced.
As a result, the value is not updated in the case LLC queue becomes
drained but last few bytes are drained at exactly that moment.
Furthermore, then the IDLE tbf timer (X2031, keep_open())) returns always
true since according to it the drain never happened.
The impact of the bug is basically delaying a bit more than expected the
time the TBF stays in IDLE state with the TBF release process yet
to be started.
Related: OS#4849
Fixes: 7d0f9a0ec3
Change-Id: I7420aeffda3500bcdc990291e4a56511af433ff9
Since a while ago, the data architecture was changed so that TBF is
guaranteed to always have a MS object associated. Hence, it makes no
sense to pass the MS object as a separate param as we can take it from
tbf object and makes code less confusing.
Change-Id: Idc0c76cf6f007afa4236480cdad0d8e99dabec5f
Before this patch, allocate_usf() was implemented to only allocate 1 USF
per TBF, regardless of the available ul_slot mask.
As a result, only 1 slot at max was allocated to any TBF. That's a pity
because usual multislot classes like 12 support up to 2 UL slots per TBF
(in common TS with DL).
This patch reworks allocate_usf() to allocate as many UL multislots as
possible (given mslot class, current USF availability, TFI availability,
related DL TBF slots for the same MS, etc.).
As a result, it can be seen that AllocTest results change substantially
and maximum concurrent TBF allocation drops under some conditions.
That happens due to more USFs being reserved (because each TBF has now
more UL slots reserved). Hence now USF exhaustion becomes the usual
limitation factor as per the number of concurrent TBFs than can be
handled per TRX (as opposed to TFIs previously).
Some of the biggest limitations in test appear though because really
high end multislot classes are used, which can consume high volumes of
UL slots (USFs), and which are probably not the most extended devices in
the field.
Moreover, in general the curren timeslot allocator for a given
multislot class will in general try to optimize the DL side gathering
most of the possible timeslots there. That means, for instance on ms
class 12 (4 Tx, 4Rx, 5 Sum), 4 DL slots and 1 UL slot will still be
selected. But in the case where only 3 PDCHs are available, then with
this new multi-slot UL support a TBF will reserve 3 DL slots and 2 UL
slots, while before this patch it would only taken 1 UL slot instead of
2.
This USF exhaustion situation can be improved in the future by
parametrizing (VTY command?) the maximum amount of UL slots that a TBF
can reserve, making for instance a default value of 2, meaning usual
classes can gather up 2 UL timelosts at a time while forbidding high-end
hungry classes to gather up to 8 UL timeslots.
Another approach would be to dynamically limit the amount of allowed
reservable UL timeslots based on current USF reservation load.
Related: OS#2282
Change-Id: Id97cc6e3b769511b591b1694549e0dac55227c43
The old VTY command is marked as deprecated and still overrides the use
in case it's used.
Related: SYS#5358
Depends: libosmocore.git Change-Id I74fb0a3afc1ac4aadbfc609b882d929401f790eb
Depends: osmo-bsc.git Change-Id I8b97ea11bad5fe05f2f634945b5703ee9abde81d
Change-Id: I46f2a955b157a409055fca7fb917dc4f75482426
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: I7f9c56313cfaa74ebe666f44763a83d8102f5484
Related: OS#5034
All vectors should be valid, since they were generated by an MS.
As can be seen, osmo-pcu fails to decode one of the vectors.
Change-Id: I37a2ddd394eeffa1cae0f3e419eeee0200a57fcf
OS#4955
Change-Id: Ib5677048f5668185ffe752f97c97d5612eee4d72
A new nacc_fsm is introduced per MS object, with its partner priv
structure struct nacc_fsm_ctx, which exists and is available in the MS
object only during the duration of the NACC procedure.
The NACC context is created on an MS whenever a Pkt Cell Change
Notification is received on Uplink RLCMAC, which asks for neighbor
information of a given ARFCN+BSIC.
First, the target ARFCN+BSIC needs to be translated into a CGI-PS
(RAC+CI) address. That's done by asking the BSC through the Neighbour
Resolution Service available in osmo-bsc using the CTRL interface.
Once the CGI-PS of the target cell is known, PCU starts a RIM RAN-INFO
request against the SGSN (which will route the request as needed), and
wait for a response containing the SI bits from the target cell.
After the SI are received, the scheduler is instructed to eventually
poll a TBF for the MS originating the CCN, so that we can send the SI
encapsulated into multiple Packet Neighbor Cell Data messages on the
downlink.
One all the SI bits are sent, the scheduler is instructed to send a
Packet Cell Change Continue message.
Once the message above has been sent, the FSM autodestroys itself.
Caches are also introduced in this patch which allows for re-using
recently known translations ARFCN+BSIC -> CGI-PS and CGI-PS -> SI_INFO
respectively.
Change-Id: Id35f40d05f3e081f32fddbf1fa34cb338db452ca
Under some circumstances, it could happen that a DL TBF is created as a
GPRS TBF due to not yet having enough information of the MS, and only
after the TBF is created the PCU gains that information and upgrades the
MS mode to "EGPRS". Hence, there's the possibility to run into a
situation where a GPRS TBF is attached to a EGPRS MS.
It may also happen sometimes that despite the TBF and the MS be EGPRS,
there's need to further limit the DL MCS to use, eg. MCS1-4 (GMSK).
As a result, when asking for the current DL (M)CS to use, we must tell
the MS which kind of limitations we want to apply. The later reasoning
was already implemented when GPRS+EGPRS multiplexing was added, but the
former was not being checked. Hence, by further spreading through the
call stack the "req_kind_mode" we match both cases.
Related: OS#4973
Change-Id: Ic0276ce045660713129f0c72f1158a3321c5977f
This patch doesn't really tests whether osmo-pcu can work on a multi-bts
environment, but it prepares the data structures to be able to do so at
any later point in time.
Change-Id: I6b10913f46c19d438c4e250a436a7446694b725a
There's no real point in using C++ there, and using C++ makes the
compiler fail to use llist_head in multi-bts patches added later due to:
"""
'offsetof' within non-standard-layout type is conditionally-supported
"""
Change-Id: I8965b5cc5a713e64788b5b6aa183d3035341ddbb
There's no real use of C++ in that file, and it causes problems when
using llist_head entry macros in future patches adding initial support
for multiple BTS in PCU object, so let's move it to plain C.
Change-Id: Ic771a89fd78b5e66151a5384f0ff6a8895589466
There's no BTS single global object anymore, get rid of those APIs. Move
users to use "pcu->bts", which will evolve to a linked list in the
future.
Change-Id: I9cf762b0d3cb9e2cc3582727e07fa82c8e183ec5
Previous work on BTS class started to get stuff out of the C++ struct
into a C struct (BTS -> struct gprs_glcmac_bts) so that some parts of
it were accessible from C code. Doing so, however, ended up being messy
too, since all code needs to be switching from one object to another,
which actually refer to the same logical component.
Let's instead rejoin the structures and make sure the struct is
accessible and usable from both C and C++ code by rewriting all methods
to be C compatible and converting 3 allocated suboject as pointers.
This way BTS can internally still use those C++ objects while providing
a clean APi to both C and C++ code.
Change-Id: I7d12c896c5ded659ca9d3bff4cf3a3fc857db9dd
Currently the BTS object (and gprs_rlcmac_bts struct) are used to hold
both PCU global fields and BTS specific fields, all mangled together.
The BTS is even accessed in lots of places by means of a singleton.
This patch introduces a new struct gprs_pcu object aimed at holding all
global state, and several fields are already moved from BTS to it. The
new object can be accessed as global variable "the_pcu", reusing and
including an already exisitng "the_pcu" global variable only used for
bssgp related purposes so far.
This is only a first step towards having a complete split global pcu and
BTS, some fields are still kept in BTS and will be moved over follow-up
smaller patches in the future (since this patch is already quite big).
So far, the code still only supports one BTS, which can be accessed
using the_pcu->bts. In the future that field will be replaced with a
list, and the BTS singletons will be removed.
The cur_fn output changes in TbfTest are actually a side effect fix,
since the singleton main_bts() now points internally to the_pcu->bts,
hence the same we allocate and assign in the test. Beforehand, "the_bts"
was allocated in the stack while main_bts() still returned an unrelated
singleton BTS object instance.
Related: OS#4935
Change-Id: I88e3c6471b80245ce3798223f1a61190f14aa840
When allocating multiple slots for a UE the following example
is not allowed 'UU----UU' for a UE class 12.
The time slot number can not roll over 7 and move to 0.
44.060 or 45.002 only specifies contigous however it was unclear
it this is an allowed pattern.
Only the example 45.002 B.3 in release 12 cleared this up.
It gives an example for a multi slot class 5 UE which has 7 possible
configuration this means the rolled over is not allowed.
Multislot class type 2 UE doesn't have this limitation.
Further if a UE supports 8 time slots this is not a limitation because
the window size (45.002 B.1) can include all time slots.
Releated: SYS#5073
Change-Id: I16019bdbe741b37b83b62749b840a3b7f4ddc6c7
When both TBFs (Dl, Ul), are detached, ms_detach_tbf() will call
ms_start_timer() which will hold a reference of the MS (ms_ref()) and
wait for X seconds (VTY config, T=-2030, 60 seconds by default) before
unrefing the MS, which will trigger ms_update_status() finally (ref==0)
and will in turn call cb.ms_idle(), which will tell the ms_storage to
free the MS.
This mechanism is used to keep MS objects around for a certain time so
that when new TBFs are established, we have cached interesting
information about the MS, ready to use.
However, in AllocTest, tons of MS are allocated in a loop calling a
function (such as test_alloc_b_ul_dl()). In that function, a BTS is
allocated in the stack and at the end of the function BTS::cleanup() is
called due to implicit destructor, which ends up calling
ms_storage::cleanup() which removes all MS from its list and frees them
*if they are not idle*. The problem here, is that due to T=-2030, an
extra reference is hold and hence the ms is not considered idle
(ms_is_idle() checks ms->ref==0). As a result, the MS is never freed,
because we don't use libosmocore mainloop here (and in any case, it
would take 60 seconds to free it).
By setting the timeout of T=-2030 to 0, ms_start_timer will avoid using
the timer and will also avoid holding the extra reference, hence
allowing ms_storage to free the object during cleanup().
This fix really helps in improving performance for AllocTest specially
after MS object contains a rate_ctr. As tons of MS objects were left
alive, they stood in the rate_ctr single per-process queue, making the
test last crazy amount of time and spending 50% of the time or more
iterating the list full of MS related rate counters.
Change-Id: I6b6ebe8903e4fe76da5e09b02b6ef28542007b6c
It is expected that the tbf object is freed at any moment in time, for
instance if osmo-pcu drops PCUIF connection with osmo-bts. I couldn't
find any reason why it would e dangerous to free the tbf, so let's
remove this message.
related: OS#4779
Change-Id: I4ab5ccaa5bf6257b18d8fd5ba06baab083821817
As we integrate osmo-pcu more and more with libosmocore features, it
becomes really hard to use them since libosmocore relies heavily on C
specific compilation features, which are not available in old C++
compilers (such as designated initializers for complex types in FSMs).
GprsMs is right now a quite simple object since initial design of
osmo-pcu made it optional and most of the logic was placed and stored
duplicated in TBF objects. However, that's changing as we introduce more
features, with the GprsMS class getting more weight. Hence, let's move
it now to be a C struct in order to be able to easily use libosmocore
features there, such as FSMs.
Some helper classes which GprsMs uses are also mostly move to C since
they are mostly structs with methods, so there's no point in having
duplicated APIs for C++ and C for such simple cases.
For some more complex classes, like (ul_,dl_)tbf, C API bindings are
added where needed so that GprsMs can use functionalitites from that
class. Most of those APIs can be kept afterwards and drop the C++ ones
since they provide no benefit in general.
Change-Id: I0b50e3367aaad9dcada76da97b438e452c8b230c
According to:
* 3GPP TS 44.060 version 16.0.0 "9.3.1a Delayed release of downlink Temporary Block Flow"
* 3GPP TS 44.064 version 16.0.0 "6.4.2.2 Unconfirmed Information (UI) Dummy command"
LLC UI Dummy frames are to be used when there no more data to send, only
in order to delay the release of a TBF. Hence, while not incorrect per
se, makes no sense to send those LLC UI Dummy frames inserted into
rlcmac blocks which already contain other LLC frames, since the MS in
that case is already being kept active.
It only makes sense to send those LLC UI Dummy frames when we have
nothing else to send, that is, alone inside a RLCMAC block without other
LLC frames.
Related: OS#4849
Change-Id: Ifae1a7b2b3dfad8df19585063088ba0df2749c8f
In previous status, if USF for GPRS-only MS was selected, then EGPRS
TBFs were skipped and either a GPRS TBF was selected or a Dummy Block
was sent. That means the behavior was unfair towards EGPRS TBFs, because
sometimes they were skipped in favor of GPRS ones.
This patch imporves the situation in the above mentioned USF scenario, by
first, under specific conditions, allowing selection of an EGPRS TBF and
then forcing it to transmit in EGPRS-GMSK (MCS1-4) so that the
USF-targeted MS can still decode the USF, while at the same time
providing more fairness by allowing the EGPRS TBF to transmit data.
The specific conditions mentioned above are, mainly, related to the fact
that once a DL data block has been sent, and hence a BSN was assigned to
it, it cannot be retransmitted later using another MCS, since lower
MCS1-4 wouldn't be able to contain higher MCS RLC payload.
The set of conditions could be expanded in the future by also selecting
the EGPRS TBF if retransmition is required and the block to be
retransmitted was originally transmitted as MCS1-4.
Related: OS#4544
Change-Id: I9af23e175435fe9ae7b0e4119ad52fcd4707b9ca
There's actually 3 errors:
* Its value should be updated, not the pointer itself
* Value should be increased, not decreased
* bitvec_read_field() API is already advancing it, no need to do it
Fixes: OS#4838
Change-Id: I009abc373794e148091e637ffee80c6461960945
The assumption that TLLI 0x00000000 is invalid and can be used
as the initializer is wrong. Similar to TMSI, 0x00000000 is a
perfectly valid value, while 0xffffffff is reserved - use it.
According to 3GPP TS 23.003, section 2.4, a TMSI/P-TMSI with
all 32 bits equal to 1 is special and shall not be allocated by
the network. The reason is that it must be stored on the SIM,
where 'ff'O represents the erased state. According to section
2.6 of the same document, a local/foreign TLLI is derived from
P-TMSI, so the same rule applies to TLLI.
I manually checked and corrected all occurances of 'tlli' in the
code. The test expectations have been adjusted with this command:
$ find tests/ -name "*.err" | xargs sed -i "s/0x00000000/0xffffffff/g"
so there should be no behavior change. The only exception is
the 'TypesTest', where TLLI 0xffffffff is being encoded and
expected in the hexdump, so I regenerated the test output.
Change-Id: Ie89fab75ecc1d8b5e238d3ff214ea7ac830b68b5
Related: OS#4844
BTS simply notifies the PCU about the supported MCS, and PCU is
responsible for providing correct data formatting supported for the BTS
and the target MS.
Related: OS#4544
Change-Id: Ifcf23771bd23afc64ca6fea38948f98f2d134ecb
For instance if PCU received DL data to be sent to an MS from an SGSN,
and the MS is not currently cached in the PCU (because there's no TBF
active for it), it will page it and transmit the DL data to it.
The SGSN is capable of sending (EGPRS) MS Class information in that same
DL data message, so it's the one responsible for providing that
information if not available at the PCU.
In the PCU if we don't have information about that MS and SGSN didn't
provide us information about it, we cannot assume the MS is going to be
EGPRS capable and even less expecting a specific EGPRS MS class.
So let's drop this code.
Related: OS#4544
Change-Id: Icce66cadb51af25ae0c3b3719940eccb548fe33b
Some tests were wrong (TypesTest) and required modification, since they
were setting a EGPRS MS but then expecting a GPRS assignment.
Change-Id: I9d3ee21c765054a36bd22352e48bde5ffca9225a
Take into account the MCS values supported by the BTS. In osmo-bts,
in general all MCS are enabled if "mode egprs" is selected in BSC,
and none otherwise.
Change-Id: Ie8f0215ba17da1e545e98bec9325c02f1e8efaea
This way everytime any program or test initiates a BTS object, the
bts_data structure has the same values.
Change-Id: Iffd6eecb1f08bda0091f45e2ef7c9c63b42e10b3
If SGSN provides us with MS class information upon DL data, let's use it
and set it in an already existing MS object if not yet known.
Also remove all unneeded code passing ms_class to append_data() which
would simply try to (again) set the ms_class.
Change-Id: I4979c9344bffd3ba7657bbab94981d233eab801f
Before this patch, it would always allocate all TBFs on the first TRX
until all TFIs were filled, then second, and so on. But it would
actually fail around 8th MS requesting an UL TBF because despite a TFI
was successfuly assigned, because all USFs were already exhausted for
that PDCH.
Related: OS#1775
Change-Id: Iccfc8acfbfdc258ed16cc5af01f12b376fe73b72
Avoid passing tons of params to internal helper function
tbf_nel_dl_assignment() in order to either fetch again the ms object or
create a new one. Let's instead create the ms function earlier if needed
and fill it with all the discovered information prior to calling the
helper function. This provides cleaner code and also better log output.
This way we also avoid trying to fill the MS twice and unneeded
getter+setter for TA.
tbf::imsi(): There' always an ms, so simply forward call to
ms()->imsi().
We can also get rid of assign_imsi, since the modified code is the only
place where it's used and there's already some code in place there to
update the MS. We instead merge it with set_imsi and keep the
duplication code to catch possible bugs from callers.
Move merge_and_clear_ms from tbf class to GprsMS, where it really
belongs.
Change-Id: Id18098bac3cff26fc4a8d2f419e21641a1f4c83b
The BTS field will be used in code paths after next patch changes,
otherwise the test fails accessing the NULL pointer.
Change-Id: I5098292bdafa9f4f70fef1a053b80a33deca722c
This is another step forward towards a more clear data model where a TBF
always has a MS object (which may be lacking some information, and at a
later point when more information is found, it may actually be a
duplicated MS object and hence one duplicate removed and the TBF moved
to the object being kept).
This helps for instance in removing duplicated information stored in
the TBF which is really per MS, like ms_class, ta, etc. Since there's
always a MS object there's no need to keep a duplicate in both classes
in case there's no MS object.
It can already be seen looking at unit test logging that this kind of
data model already provides better information.
Some unit test parts were needed to adapt to the new model too.
Change-Id: I3cdf4d53e222777d5a2bf4c5aad3a7414105f14c
Return an interface to the window base class so that the tbf base class
can access the common window methods, such as set_ws(). It will be used
in next commit to get rid of duplicated function enable_egprs in both
dl_tbf and ul_tbf subclasses.
The user of the function can then decide to access more specific
functionaltiites of the window class by static casting it to the
specific direction (which is known by the caller since it operates on a
ul_tbf or a dl_tbf).
Change-Id: Ia2e1decf91be1184668e28297c2126affb9c7ae4
In those cases since a string pointer was passed, it always printed
"single" instead of whatever really was being used, since the string
pointer was not NULL.
Change-Id: Idab7d18e8f519e10fc3df4007634661c46f9256d
In general we want to see explicitly the kind of requested allocation at
the start. The MS class is not needed since it's printed in the previous
log line in any case.
Change-Id: I9eb0a592c15be96da9d140ff373c1afead76b18c
Add support of the second NSVC in the info indication.
Add support to update a previous NS configuration.
Allow to update of a NS-VC while the NSE is still available over the
second.
Depends-on: I917f25ebd1239eae5855d973ced15b93731e33a0 (libosmocore)
Depends-on: I3a0cd305fd73b3cb9ec70246ec15ac70b83e57f2 (libosmocore)
Depends-on: I5a2bb95d05d06d909347e2fb084a446ead888cb3 (libosmocore)
Depends-on: I54f110acc3acccb362f6e554324d08cc42b7c328 (libosmocore)
Depends-on: Ia00753a64b7622a0864341f51ea49b6963543755 (libosmocore)
Depends-on: Ic8f6f8aca10da23a18fab8870be7806065a34b47 (libosmocore)
Depends-on: I5f67e6a9bf4cb322bd169061fee0a528012ed54d (libosmocore)
Change-Id: I589ebaa2a2b7de55b7e4e975d8fd6412dd5f214b
Pau Espin
Oliver Smith
Neels Hofmeyr
Vadim Yanitskiy
Harald Welte