This one is larger than some of the other already available.
The decoder is wroking as expected here.
Change-Id: I5d986f68395326f894349446194090b1ddaecd69
Fixes following error log line:
"{RELEASING}: Event DL_ACKNACK_MISS not permitted"
Rationale: We may move to RELEASING state at some point, for instance
due to MAX_N3101/MAX_N3105 while still having some active poll
registered in some PDCH ulc. Upon that poll (most probably) timing out,
it will send a DL_ACKNACK_MISS event to us. Since we are already
determined to release the TBF (waiting for T3195 or T3169 to trigger),
simply ignore the event and avoid logging an error.
Fixes: OS#5240
Change-Id: Ibfb49356d2b3b5fccb6d59db8593b2256e5c51fb
This clarifies the different paths and uniforms them. Makes code far
easier to read and debug.
New improved verification already found some misehavior in some tests.
Change-Id: I7e4a88d6e004bbb7974595320ed73742162c7ad7
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
Move FSM internal state checks to its own file. Re-use the helper
function in the 2 places where same stuff is checked.
Change-Id: I9ded6e1c80e6cd7bcf6883bc2e853b6dafb33f7c
As seen operating PCU after BTS restart, lots of following message
sequences due to FSM kept in same state (hence scheduler retyring every
time):
"""
DTBF tbf_ul_ass_fsm.c:306 UL_ASS_TBF(DL-TFI_0){SEND_ASS}: Received Event CREATE_RLCMAC_MSG
DTBF tbf_ul_ass_fsm.c:95 TBF(TFI=0 TLLI=0xf80bd801 DIR=DL STATE=RELEASING EGPRS) We have a schedule for uplink assignment, but there is no uplink TBF
DTBF tbf_ul_ass_fsm.c:97 UL_ASS_TBF(DL-TFI_0){SEND_ASS}: transition to state NONE not permitted!
DTBF tbf_ul_ass_fsm.c:306 UL_ASS_TBF(DL-TFI_0){SEND_ASS}: Received Event CREATE_RLCMAC_MSG
"""
Change-Id: I91d74f70a9106ccbf0c137b6e713877f9ea8f59d
Make use of the separate GPRS counters added in previous patch
I0c0a1121b4ae5f031782e7e63a0c28eb0b6c8b42 to shorten
has_gprs_only_tb_attached.
Related: SYS#4878
Change-Id: I1dd7df2c740ea604f07c65bebcb7c0051aebf9ae
Same was already done for PDTCH in previous commits. Let's now apply
same bits to PTCCH.
Related: SYS#4919
Change-Id: If6617964e67fc35eeee1791b06e13bf63ac88f73
We also want to avoid sending idle blocs in TRX0 to the BTS, so that the
BTS can be aware of blocks being idle and then submitting dummy blokcs
by itself applying required BCCH Carrier power reduction.
Related: SYS#4919
Change-Id: Idd58d2a09c3947098b960cfcb5cd1b7b7bca3d84
Add stats needed for performance measurements in
3GPP TS 52.402 § B.2.1.54-55.
Split m_num_tbfs to count GPRS and EGPRS TBFs separately. Move the code
that updates m_num_tbfs and sets the PDCH_OCCUPIED stats to a separate
function, as it's mostly the same in the TBF attach and detach.
Related: SYS#4878
Change-Id: I0c0a1121b4ae5f031782e7e63a0c28eb0b6c8b42
We already have a similar function for Neighbor Address Resolution.
This way we keep as much as possible internal state related logic into
the nacc_fsm.c file.
Change-Id: I7378939825cc3ec3280f76bc51233c0a172d8a27
While NACC was initially developed, it became clear there was need for
a way to interact PCU<->BSC in order resolve ARFCN+BSIC into CGI-PS
for later RIM usage.
Hence, this resolution was first (until today) implemented using an out
of bands RPC system using the CTRL interface, which required specific
config to be written and matches in osmo-pcu and osmo-bsc VTY (ip+port
of the CTRL interface to use).
However, this has several shortcomings:
* As explained above, specific configuration is required
* Since recently, we do support BSC redundancy in osmo-bts. Hence the BTS
may switch to a BSC other than first one. If that happened, that'd mean
the CTRL interface would still point to the initially configured one,
which may not be the same currently serving the PCU.
During recent development of ANR related features, a similar need for
PCU<->BSC was required, but this time it was decided to extend the IPA
multiplex of the Abis OML connection to pass PCUIF messages,
transparently forwarded to each side by the BTS.
This has the advantage that connection PCU<->BTS is handled by BTS and
both sides send messages transparently.
Let's switch by default to using this new interface, while still
maintaing the old way for a while (announcing them as deprecated) to
avoid breaking existing deployments until they are upgraded to new
versions of osmo-pcu and osmo-bsc.
Related: SYS#4971
Change-Id: I6ad33c7ab10202840cf804dea9ba595978d0e920
Count available PDCHs (3GPP TS 52.402 § B.2.1.38) as well as occupied
PDCHs (§ B.2.1.42-44).
Related: SYS#4878
Change-Id: I74760a68ee055510a79e80854ec7bf1521669119
When a PDCH TS becomes disabled (eg due to dyn TS being used for a
call), we are currently freeing all attached PDCHs in order to avoid
further use of it. However, pdch_free_all_tbf() was only freeing TBFs
attached to the PDCH, that is, TBFs having a valid TFI assigned.
There are some cases where temporary dummy TBFs are created which have
no TFI assigned, such as when creating an ImmAssReject. Let's take those
into account too, and make sure they are freed.
Related: OS#5226
Change-Id: Ibfe78448ebdedc8b049c80664711e166d910f9b7
These messages are expected under some circumstances, such as when
direct phy is used and a chan is disabled (eg. dyn TS). This happens
because PCU asks for chan de-activation through PCUIF while at the same
time marking the PDCH locally as disabled. Hence, in the time the BTS
manages to disable it on the lower layers, the phy still sends us
RTS indications.
Let's keep it under NOTICE to avoid clogging the logs in production
setups which are usually using global level of NOTICE or ERROR.
Related: OS#5222
Change-Id: Iab9e1590b504bf05dc693e27550b30db0dffcbc7
It seems there may be a race conditon where lower layers (direct PCU)
send UL blocks to us while the PDCH was already disabled (due to a call
entering on a dynamic TS).
As the PDCH is disabled, the ULC is NULL and shouldn't be used before
being enabled again.
Related: OS#5222
Change-Id: I4b8931f0cc7cfc787a1cc35196295402524b15c3
When setting a POLL, it will always happen on PACCH, so all the CCCH
part makes no sense there. Let's drop it and move the logging of each
case to the caller, where logging file+line is more useful.
Change-Id: I242f97fd6f927131ac64c1a7c9c3812b6389de04
Method is renamed since it clearly relates to getting DL ACK/NACK, no
CTRL ACK.
use same methods in both scheduler and internal use since they are
expectd to be run in the same code path by the scheduler. This way we
make sure the same conditions apply and it's clearer when looking at
the code.
Change-Id: Ib0e9b9547f5292b95064bab2dc182fdf659f0518
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 never use the std:string anyway, we always call .c_str() to log using
osmocom logging system.
Furthermore, we'll need to use it from C code soon (next commit).
Change-Id: I3ad66f9f3f4d55d11da3a3b8b38656ae2dd50603
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
Having that code in a separate function is confusing and adds code
complexity since it looks like an entry point to start feeing a TBF, but
it simply some (not yet really useful) set of instructions to be called
one 1 code path in tbf_free.
Let's move it there, this way it becomes clear tbf_free() is THE place
to be (if you want to get rid of a TBF).
Change-Id: I30febf4d21a0bfab37524c07598bbb0dd32f7f65
This way we clean up tbf_free entry point, and leave memory freeing for
later on at the end when talloc_free is called.
Change-Id: I1c45e3296e565725bcbbca391d9518772fffa89d
Function is already called by gprs_rlcmac_received_lost(), so next call
following it will be sum=0 and return EINVAL.
Change-Id: I015ba16d18fdd6e2441ec3c256b5ac88771d7a8b
