OpenBSC March 2018

openbsc@lists.osmocom.org

16 participants
35 discussions

New (but unofficial) asn1c has been applied to osmo-iuh
by Chiu Bi-Ruei 29 Mar '18

29 Mar '18

Dear All, I have just uploaded my modified osmo-iuh, which a new version asn1c is applied, to : https://github.com/brchiu/osmo-iuh/tree/new-asn1c Code generated by this new asn1c can handle necessary ASN.1 information object class feature used in several 3gpp specs, thus it remove the need of asn1tostruct.py. Because Lev Walkin has not accepted corresponding pull requests yet, it should be built from Velchkov's repository : https://github.com/velichkov/asn1c/tree/s1ap And link to new libasn1c : https://github.com/brchiu/libasn1c/tree/new-asn1c I use RANAP 14.1.0, RUA 14.0.0 and HNBAP 14.0.0 to generate files. There are still several compilation error which are irrelevant to this new asn1c. https://gist.github.com/brchiu/17b5d306dd3e95f5e7b255a7dbb81344 Hope someone can help solving them. Because I do not have test environment nor have no time to verify this porting, you are welcomed to use this result on your own. thanks. regards, Bi-Ruei, Chiu.

2 1

git.osmocom.org read-only during migration
by Harald Welte 28 Mar '18

28 Mar '18

Hi all, I've set git.osmocom.org to read-only while I'm doing the migration to the new server. So don't be surprised if you won't be able to push there (or if gerrit is not able to push there). I'll try to make this quick, should be back within the afternoon. -- - Harald Welte <laforge(a)gnumonks.org> http://laforge.gnumonks.org/ ============================================================================ "Privacy in residential applications is a desirable marketing option." (ETSI EN 300 175-7 Ch. A6)

1 2

general review of FSMs across the code base
by Neels Hofmeyr 26 Mar '18

26 Mar '18

When libosmocore/contrib/fsm-to-dot.py hinted me at errors in the new gscon FSM in osmo-bsc, I took a moment to render and look at all our FSM definitions. The osmo-bsc ones are fixed by https://gerrit.osmocom.org/7501 I also found OS#3110 and OS#3111 I published the FSM graphs as rendered today at http://people.osmocom.org/neels/osmo_fsm_graphs/ and might do so every now and then. If anyone likes, we could do it from jenkins regularly. ~N

1 0

"lai_and_lac"
by Neels Hofmeyr 26 Mar '18

26 Mar '18

Hi stsp, I noticed in the recent gsm0808_cell_id_list2 a review item slipped by: A "LAI" is by definition a PLMN (MCC+MNC) plus a LAC. So saying "LAI and LAC" is like saying "My family and my brother". So maybe we want to rename the "lai_and_lac" member of struct gsm0808_cell_id_list2 before it gets tagged for release. Just "lai" would be accurate, as its type (struct osmo_location_area_id) suggests. ~N

2 2

OSMO-BTS-TRX: Cannot send payload with invalid length! (expecting 15, received 14)
by Ron 23 Mar '18

23 Mar '18

Hi Support, In Half Rate configuration of OSMO-BSC, is there a way to used ETSI TS 101 318 standard for its RTP Payload Format? As per testing, the RTP Payload format used by OSMO is RFC5993. Can this be configured in OSMO-BSC or OSMO-BTS-TRX? Or the configuration of this is hard coded? Logs from OSMO-BTS-TRX: [0;m<0015> trau/osmo_ortp.c:0 Receiving packet with unknown payload type 111.[0;m<0006> scheduler_trx.c:543 Cannot send payload with invalid length! (expecting 15, received 14) [0;m<0015> trau/osmo_ortp.c:0 Receiving packet with unknown payload type 111.[0;m<0006> scheduler_trx.c:543 Cannot send payload with invalid length! (expecting 15, received 14) [0;m<0015> trau/osmo_ortp.c:0 Receiving packet with unknown payload type 111.[0;m<0006> scheduler_trx.c:543 Cannot send payload with invalid length! (expecting 15, received 14) [0;m<0006> scheduler_trx.c:543 Cannot send payload with invalid length! (expecting 15, received 14) [0;m<0006> scheduler_trx.c:543 Cannot send payload with invalid length! (expecting 15, received 14) [0;m<0006> scheduler_trx.c:543 Cannot send payload with invalid length! (expecting 15, received 14) [0;m<0006> scheduler_trx.c:543 Cannot send payload with invalid length! (expecting 15, received 14) [0;m<0006> scheduler_trx.c:543 Cannot send payload with invalid length! (expecting 15, received 14) [0;m<0006> scheduler_trx.c:543 Cannot send payload with invalid length! (expecting 15, received 14) [0;m<0015> trau/osmo_ortp.c:0 Receiving packet with unknown payload type 111.[0;m<0006> scheduler_trx.c:543 Cannot send payload with invalid length! (expecting 15, received 14) [0;m<0015> trau/osmo_ortp.c:0 Receiving packet with unknown payload type 111.[0;m<0006> scheduler_trx.c:543 Cannot send payload with invalid length! (expecting 15, received 14) [0;m<0015> trau/osmo_ortp.c:0 Receiving packet with unknown payload type 111.[0;m<0015> trau/osmo_ortp.c:0 Receiving packet with unknown payload type 111.[0;m<0015> trau/osmo_ortp.c:0 Receiving packet with unknown payload type 111.[0;m<0015> trau/osmo_ortp.c:0 Receiving packet with unknown payload type 111.[0;m<0015> trau/osmo_ortp.c:0 Receiving packet with unknown payload type 111.[0;m<0006> scheduler_trx.c:543 Cannot send payload with invalid length! (expecting 15, received 14) [0;m<0015> trau/osmo_ortp.c:0 Receiving packet with unknown payload type 111.[0;m<0015> trau/osmo_ortp.c:0 Receiving packet with unknown payload type 111.[0;m<0015> trau/osmo_ortp.c:0 Receiving packet with unknown payload type 111.[0;m<0015> trau/osmo_ortp.c:0 Receiving packet with unknown payload type 111.[0;m<0006> scheduler_trx.c:543 Cannot send payload with invalid length! (expecting 15, received 14) TIA! Best Regard, Ron Menez ron.menez(a)entropysolution.com<mailto:ron.menez@entropysolution.com>

2 2

MSC subscriber conn FSM issues
by Neels Hofmeyr 19 Mar '18

19 Mar '18

In recent weeks, many of you have encountered and worked at failures of OsmoMSC's subscriber conn FSM / the conn itself to handle corner cases and/or cleanly handle error situations. I have plans for that FSM to be reviewed, adding probably two more states to it, a) to properly handle the early startup and b) gracefully release in all situations; and to probably more closely glue the conn struct to the FSM. My humble suggestion is to wait for me to review it and not spend more time on it until that is through. I wrote the FSM initially, and it makes sense that I clarify its structuring before all of you need to learn its current quirks. I hope it won't take too long, just need to sort the priorities of this fix versus the other tasks I have... ~N

1 0

changes in osmo-ttcn3-hacks (logging, external dependencies)
by Harald Welte 19 Mar '18

19 Mar '18

Dear all, I've changed some details related to logging: * we now generate per-testcase merged TITAN log files with the same name prefix as the pcap files (Module.Testcase.{pcap,merged}) and remove the hundreds of per-component log files generated by the parallel executor after merging them. * the console log verbosity has been incresed to include all output from explicit "log()" statements. This is particularly useful when running the tests interactively. Unrelated to the above, I removed our own copies of the MTP3/M3UA/SCCP emulation and replaced it with the proper upstream git repositories under 'deps' like most other external modules. This means that you will need to * update your 'deps' (make deps-update) * remove/re-generate your symlinks (gen_links.sh) from e.g. 'bsc' or 'msc' * re-generate the makefile subsequently It may be that the make file dependency logic doesn't pick this up automatically, so if you see some strange build errors, it may be a good idea to clean your local work tree. Regards, Harald -- - Harald Welte <laforge(a)gnumonks.org> http://laforge.gnumonks.org/ ============================================================================ "Privacy in residential applications is a desirable marketing option." (ETSI EN 300 175-7 Ch. A6)

2 1

BSC_Tests.ttcn now needs TITAN 6.3.1 to build
by Harald Welte 16 Mar '18

16 Mar '18

Dear all, as I explained in https://osmocom.org/issues/3070 for some reason BSC_Tests.ttcn fails to work under TITAN 6.1.0 anymore. Rather than debugging old TITAN versions, I decided to update to the latest stable release (6.3.1) in the debian-stretch-titan container that we use on jenkins. Please keep this in mind when running tests locally: * for containerized setups, switch to latest docker-playground master and re-build the debain-stretch-titan container before building any of the ttcn3-*-test containers based on it * if you build+run the ttcn-3 code on your native machine/distro, pleaes switch to TITAN 6.3.1. If you run Debian 9, you can use the package from network:osmocom:latest Regards, Harald -- - Harald Welte <laforge(a)gnumonks.org> http://laforge.gnumonks.org/ ============================================================================ "Privacy in residential applications is a desirable marketing option." (ETSI EN 300 175-7 Ch. A6)

1 0

docker-playground to gerrit?
by Neels Hofmeyr 16 Mar '18

16 Mar '18

Do I have permission to move the docker-playground.git to gerrit? I have a bunch of patches already in my various branches. I could just push them onto master, but I guess some code review would be good? Should we also add a gerrit job that tests the changes? Kind of hard to make a failure decision though while when some tests are normally failing, like they still are. ~N

2 1

ttcn3: BSSGP_Types.ttcn MCC-MNC
by Neels Hofmeyr 16 Mar '18

16 Mar '18

In the file BSSGP_Types.ttcn -> ../deps/titan.ProtocolModules.BSSGP_v13.0.0/src/BSSGP_Types.ttcn I notice many PLMN definitions like type record PLMN_Identity { OCT1 iEI, BIT1 ext, LIN2_2a lengthIndicator, HEX1 mccDigit1, HEX1 mccDigit2, HEX1 mccDigit3, HEX1 mncDigit3, HEX1 mncDigit1, HEX1 mncDigit2 } which at first glance looks like they got the MCC-MNC digits ordered wrongly. It is correct as long as the less significant nibble comes first. And using these in PLMN tests gives the expected results. Now I assume that the HEX1 means that it's a nibble, where the less significant nibble comes first, sort of a "network nibble order" if that makes any sense. It is weird, though -- do we need to compose hex strings "reversed" as well?? ~N

2 1

2025

2024

2023

2022

2021

2020

2019

2018

2017

2016

2015

2014

2013

2012

2011

2010

2009

OpenBSC March 2018