June 2011 - OpenBSC - lists.osmocom.org

GPRS - local and foreign TLLIs
by Luca Melette 04 Feb '26

04 Feb '26

Hello everybody, after setting up a nanoBTS with OpenBSC/SGSN/GGSN, I had some troubles trying to connect my smartphone to the GPRS cell. Investigating the BTS-to-SGSN traffic, I saw that the frames sent by the SGSN were all marked with the same N(U) value (at LLC layer), the value was 0. With some debug, I found that the there was a mismatch in the TLLI storage, used to keep the status of attached terminals. The mentioned code is contained in gprs_llc.c, involving LL Entity functions and gprs_llc_tx_ui(). /* look-up or create the LL Entity for this (TLLI, SAPI) tuple */ lle = lle_by_tlli_sapi(msgb_tlli(msg), sapi); if (!lle) { struct gprs_llc_llme *llme; LOGP(DLLC, LOGL_ERROR, "LLC TX: unknown TLLI 0x%08x, " "creating LLME on the fly\n", msgb_tlli(msg)); llme = llme_alloc(msgb_tlli(msg)); lle = &llme->lle[sapi]; } The TX function uses the previously received TLLI to lookup for the LLE. The lle_by_tlli_sapi() performs the search, but preliminary it applies a foreign2local TLLI conversion to ensure to have a local one. Since no valid entry for the searched TLLI is found, a new one is created. And the problem is here. The new entry has the foreign TLLI. Next time the lookup fails again, and a new entry is created. This way, the counter N(U) is always reset, and the mobile do not recognize the message sequence and no attach is possible. My question is about foreign and local TLLIs. I patched the lookup, avoiding the conversion, so that the LLE is found and everything works fine... but... What is the sense of the conversion? Should the TLLI be always stored as a local one? Can this problem be solved with another foreign2local while allocating new entries? Actually, my problem has been solved with that workaround. But I'm curious to know what is the right way. Hope somebody can answer :) Thanks. Cheers, LM

5 6

Re: Help with sccp library
by mosbah abdelkader 07 Jul '25

07 Jul '25

> I don't understand. This callback will be called with data you need to write > to the network. In case of MTP Level3 you will need to wrap that around the > msgb you got. I means: is the interaction with mtp3 layer implemented (is sending sccp data by mtp3 implemented by the library?)? Also, what about the reception of data from mtp3 layer. is that implemented in the sccp lib. I am asking these questions because I see the code of mtp3 in the lib but no significant call is present in the sccp part of the lib. Thank you for your help.

4 3

Several questions
by jason mercury 06 Dec '12

06 Dec '12

I have some questions: 1) When I start bsc_hack bsc_init.c first establishes OML link and initializes the bts then it establishes RSL link and bts starts broadcasting. However, it takes so much time to start the bts. Instead of this I want to do the following: it establishes OML link at the beginning and only once, then when i want to start broadcasting it establishes just the RSL link and bts will start faster since i don't have to wait for OML link. What should be done for this? 2) If i send one or two word messages from telnet interface it is okay. But if i send a longer message the phone could't receive the end of the message correctly(last words may be incomplete). Did any one encounter with this problem? What is wrong with me? 3) Could I send SMS in which extension of the sender is text not integer. For example, i want to send an information SMS that this is a test network. For this purpose i want to send an SMS from 'OpenBSC'. I set the extension of the first subscriber in database as text and tried to send the SMS but SMS wasn't delivered. What should i do? 4) Can i add SMS externally to SMS table of database? Thanks. Jason

4 3

GSM @ Camp 2011 preparations
by Harald Welte 02 Aug '11

02 Aug '11

Hi Peter and others, it seems like the situation regarding the GSM network at the Camp is starting to become clearer every day. I am still arguing that two relatively large BTS with circular antennas are a good idea. Therefore, I have now applied for a test license with the regulatory authority with the following parameters: * six GSM 1800 ARFCN * two antennas, circular, 18 meters above ground * 5W output power on each ARFCN. I know one of my customers in Germany has once managed to get a license for 5W, so I thought it is a good bet we should get the same. It should be more than sufficient to cover the camp. * from August 2nd through August 15th, i.e. we have time for build-up and can also run it one more day during shutdown of the camp I should be receiving "Motorola Horizon Macro" BTSs with 3 to 6 TRX each soon. They are able to drive something like >= 20W output power, way more than we need. The only problem with those BTS is: We will need to add the Motorola A-bis dialect to OpenBSC before the Camp, which could be a tough task depending on how far they are off the standard A-bis as specified in 08.58 and 12.21. If we cannot make those BTS work, our options are somewhat limited. 1) nanoBTS based The nanoBTS are 200mW only, so we would need a booster. Those cost about 1000 EUR for 6W downlink power + 18dB uplink LNA. Who is going to fund that? Also, we would need a combiner/splitter to run something like a 3-TRX setup. 2) Ericsson RBS I should have two RBS 2308 (4TRX each) soon, but both of them are GSM 1900. It is unlikely that we ge a license, and a lot of phones probably don't support it. So unless they can somehow be converted to 1800 MHz units, they are not a good fit either :/ 3) Siemens BS-11 All of them are GSM 900. No way to get a license in Germany, sorry. So we basically _have to_ make the Motorola Horizon work... 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)

4 12

sms bugfix
by Dennis Wehrle 25 Jul '11

25 Jul '11

Hello Our goal was to send status sms via the vty interface. But all of our sms were cropped. In contrast sms from one MS to another MS are displayed correctly (despite the fact, that the text at the database contains several '@' at the end / the user_data contains several zero-octets). Therefore i have inspect the code and found several bugs. The main problem is that the "user_data_len" is not correctly used. As per GSM 03.40, 9.2.3.16 TP‑User‑Data‑Length (TP‑UDL): "If the TP‑User‑Data is coded using the GSM 7 bit default alphabet, the TP‑User‑Data‑Length field gives an integer representation of the number of septets within the TP‑User‑Data field to follow." Currently the "user_data_len" contains the number of octets (returned from gsm_7bit_encode(...) at gsm_utils.c (libosmocore)). The big problem here is that this information is not unique, e.g.: 1.) 46 non-extension characters + 1 extension character => (46 * 7 bit + (1 * (2 * 7 bit))) / 8 bit = 42 octets 2.) 47 non-extension characters => (47 * 7 bit) / 8 bit = 41,125 = 42 octets 3.) 48 non-extension characters => (48 * 7 bit) / 8 bit = 42 octects But the MS has to know the correct "user_data_len" to decode the correct number of characters. For this reason i updated the gsm_7bit_encode() function to return the correct number of septets. However sometimes it is needed to know the correct number of octets (e.g. at gsm_04_11.c: gsm340_gen_tpdu(...)) => i added a function to gsm_utils.c named: uint8_t get_octet_len(const uint8_t sept_len) I have also fixed the problem, that the sms are wrongly stored / displayed on the database. But the solution on the function *sms_from_result(...) (at db.c) is not really "beautiful". This is because there exists no "user_data_len" field at the database. To store the right value for "user_data_len" (which is further needed) i have to get the length from the "text" field. Unfortunately this is not enough. If the text contains extension characters like {[]} etc. then the "user_data_len" has to be bigger because these characters needs two septets. Therefore i use a switch statement so search for these characters. A better solution for that is to store the right "user_data_len" to the database (on the encoding / decoding procedure). But i don't know if this is a suitable solution for all of you (because you have to change your database structure etc.). Best Regards Dennis Wehrle

6 24

Licensing of the wiki
by Harald Welte 13 Jul '11

13 Jul '11

Hi all, so far we do not have any statement about the licensing of the content on our wiki. This means the default copyright rules apply: All the content is copyrighted, and nobody has any rgiht to reproduce it at all. I would like to propose an official license for the content in the wiki: Create Commons CC-BY-SA. The alternative would be to go for CC-BY-NC-SA, disallowing commercial use of the content. I'm not certain if NC is really what we want. After all, even somebody using OpenBSC in a commercial environment should be able to make copies of the reference documentation we have available - as long as he will releas the result again (which the SA part already covers). If the major contributors to the wiki would agree to a license, I would update the wiki accordingly. Thanks, 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)

3 3

[PATCH 0/4] use standalone libtalloc instead of own copy
by pablo＠gnumonks.org 05 Jul '11

05 Jul '11

From: Pablo Neira Ayuso <pablo(a)gnumonks.org> Hi! With these patches, libosmocore, libosmo-sccp, openbsc and osmo-tetra does not use anymore our own copy of talloc and they rely on the standalone library that is provided by major distributors. The osmocom-bb part is still missing, my idea is to include a copy of libtalloc in the tree in the shared/ directory via rsync. This patch will follow later since we can still rely on the outdated copy of libosmocore. Please, apply.

2 5

libosmocore / logging from within libraries
by Harald Welte 27 Jun '11

27 Jun '11

Hi! As more and more code is moving into libraries (like I just did with the LAPDm code, and like pablo is working on with libosmo-abis), we needed a solution how to allocate and use the LOGP subsystem constants like DRSL, DRR, ... from within libraries. The existing logging code wasn't really prepared for that. I've now come um with a hack to extend it while preserving compatibility to applications: * we use negative numbers starting from -1 for library-internal subsystems * those numbers get converted to a positive index into the various arrays at run-time. So -1 ends up one entry higher in the array than the last application-providede log category/subsystem. As part of this change, the array allocations are now dynamic, i.e there is no maximum limit for the number of log categories that an application can register with the core. Only for libraries (even outside libosmocore), we have compile-time registration, i.e. the 'struct log_info_cat' and the D* constant need to be defined inside libosmocore. I think this is an acceptable compromise. Furthermore, if LOGP()/DEBUGP() ever see a subsystem number that it doesn't know, it will assign it to the new 'DLGLOBAL' (Debug Library GLOBAL) category, i.e. there cna be no array overflows. This ensures that even an external library using a 'newer' D* constant will not crash or otherwise fail, it will simply log in a slightly different way. 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

Complete test of a PDP Context activation process
by Guido Cilia 22 Jun '11

22 Jun '11

Hi everybody! I need to replicate in my private network the functionalities of an SGSN and a GGSN for academic purpose. I see the openGGSN project and I understand that it's composed by two components: sgsnemu and ggsn. What I want to ask to you is if I could test all the PDP Context activation process with these two components and how. This is the real (very semplified) process: Mobile Phone ---> SGSN ---> GGSN ---> SGSN ---> Mobile Phone (for a full description of the process see here<http://www.eventhelix.com/RealtimeMantra/Telecom/gprs_attach_pdp_sequence_d…> ) My suspicion, tell me if I'm wrong, is that sgsnemu component allow to test only this part of the complete process: SGSN ---> GGSN ---> SGSN (simulating the mobile terminal request but not really accepting it from an external module/device). Is it right? How could I test all the process? Consider that I could use a linux pc instead of a mobile phone. Thank you very much!!! -- Guido Cilia Web: http://www.guidocilia.it E-Mail: guidocx84(a)gmail.com Skype: guidocx84

2 1

Motorola Mo-bis / Implications for CCC Camp
by Harald Welte 19 Jun '11

19 Jun '11

Hi all, it seems like supporting the Motorola BTS from OpenBSC is a really big challenge. Not only are there problems regarding an unknown configuration database format (see http://laforge.gnumonks.org/weblog/2011/06/18/) Dieter has found the following excellent document, which in Chapter 5 describes the architecture of the Motorola specific changes to A-bis: http://read.pudn.com/downloads61/ebook/212957/SYS01.pdf Judging from the document: * there is an unknown 'motorola executive header' in front of RSL * many RSL standard messages have been replaced by proprietary ones * they use a mixture of 08.08 (A) and 08.58 (A-bis) information elements * many features normally found in the BSC are implemented in the BTS, e.g. the entire CHAN RQD / RSL CHAN ACT / IMM ASS CMD sequence is done fully inside the BTS. As a result, I think it will be a quite complex project to support those BTSs from OpenBSC. It's not just a matter of adding some small vendor-specific OML messages like with Siemens, Ericsson and ip.access. While it might still be interesting, I think it's unrealistic to assume this would happen ahead of the camp [unless somebody finds a compatible Motorola BSC that they can borrow us for analysis + protocol tracing]. Thus, the focus for the 2011 CCC Camp is now back to ip.access. The main issues we will be facing is on the RF side, i.e. * RF PA for the TX side * combining the PA outputs to be transmitted on one antenna * filter + LNA for Rx side * splitting the Rx signal As indicated previously, I would love to see something like two BTS with three TRX each. If we'd recycle the duplexer/combiner blocks from the Motorola BTS, we could do something like three sites with 2 TRX each. (as they can only combine two TRX). We _might_ also be able to re-cycle the TX PA from the Motorola CTUs, at least on the RF side it would be pretty easy to drive them from a nanoBTS rather than its internal source (as there's a connector). The bigger question is how do we get the controller logic of the PA into a state that the PA is actually enabled and active... 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