OpenBSC

openbsc@lists.osmocom.org

6 participants
4274 discussions

nanoBTS - First Call Ok, but not others -> I have an ugly hack to fix it
by Sylvain Munaut 27 Aug '09

27 Aug '09

Hi, I had the exact same problem has described here http://lists.gnumonks.org/pipermail/openbsc/2009-July/000627.html That is, the first call worked great but after the first call, no way to make a second, the BTS would respond with NACK to the CHANnel ACTIvation requests. I finally found that the chan_rate field of the channel_mode must be 0 for that request to work. And forcing it to 0 in rsl_chan_activate_lchan, like this : --- SNIP --- rc = channel_mode_from_lchan(&cm, lchan); if (rc < 0) return rc; cm.chan_rate = 0; // Hack ! ----------- makes the thing work. But that's obviously not the root cause. I thing the lchan tch_mode should be GSM48_CMODE_SIGN, so that channel_mode_from_lchan would set the proper value directly. But I haven't traced yet why tch_mode is not at the good value for the second call ... Sylvain

2 1

nanoBTS - Different version ?
by Sylvain Munaut 27 Aug '09

27 Aug '09

Hi, I just got hold of a ip.access nanoBTS model 139U, but so far my attemps at making it work with OpenBSC have been unsuccessful :( It's a 'used' one so there are config already done most likely and I don't know which. What differs from http://bs11-abis.gnumonks.org/trac/wiki/nanoBTS : * It doesn't request an IP by DHCP It seems to have IP/Netmask and GW hardcoded. * When plugging it, it tries to do an ARP request for what I guess his the GW address, and when it get a response, it tries to connect to TCP port 3022 (yes, it's 3022 and not 3002) of yet another IP in another subnet (which I guess is the configured Primarly OML IP) * ipaccess-find does find it : ---- ipaccess-find (C) 2009 by Harald Welte This is FREE SOFTWARE with ABSOLUTELY NO WARRANTY Trying to find ip.access BTS by broadcast UDP... MAC Address='00:02:95:00:31:22' IP Address='10.105.254.142' Unit ID='65143/0/0' Location 1='NanoBTS FXMarketspace' Location 2='BTS_NBT131G' Equipment Version='139_029_41' Software Version='120a352_v267b22d0' Unit Name='69419' Serial Number='00071582' ---- * ipaccess-config fails, the nanoBTS refuses incoming connections on port tcp 3006 * Here's the result of a NMAP : ---- sh$ sudo nmap -sS 10.105.254.142 -p 1-65535 Starting Nmap 4.76 ( http://nmap.org ) at 2009-08-26 16:21 CEST Interesting ports on 10.105.254.142: Not shown: 65533 closed ports PORT STATE SERVICE 3026/tcp open unknown 3034/tcp open unknown MAC Address: 00:02:95:00:31:22 (IP.Access Limited) ---- * I then tried to run ipaccess-config to port 3026 but it fails (just waits a while thensays short read / BTS disappeared) * I tried making openbsc listen on 3022 but with no real results : --- sh$ bsc_hack DB: Database initialized. DB: Database prepared. no matching signalling link for hh->proto=0x01 no matching signalling link for hh->proto=0x29 no matching signalling link for hh->proto=0x2b no matching signalling link for hh->proto=0x1b short read! no matching signalling link for hh->proto=0x01 short read! BTS disappeared, dead socket short read! no matching signalling link for hh->proto=0x01 no matching signalling link for hh->proto=0x29 no matching signalling link for hh->proto=0x2b no matching signalling link for hh->proto=0xc9 short read! no matching signalling link for hh->proto=0x01 short read! BTS disappeared, dead socket --- Any idea any one ? Sylvain

2 2

How to tell MS no channel is available?
by Nordin 25 Aug '09

25 Aug '09

Hello guys, Suppose I have 8 MS registered on my nanoBTS and the ninth wants also to register. How can I tell the ninth MS "Sorry dude, there is no room for you, try another ARFCN"? I've checked the REJECT causes in gsm 04.08, but couldn't really find one which is informing the BTS is out of free channels. Now my mobile keeps trying to register to my nanoBTS, although I rejected him with a standard cause. Maybe someone can give me a hint where to look at? Thanks.

2 2

Cell Broadcast Support
by Galbayar D 25 Aug '09

25 Aug '09

Hi Is OpenBSC supports Cell Broadcast? If not is there any plan? or Is someone developing? Gala

3 2

How was HAR2009?
by Nordin 24 Aug '09

24 Aug '09

Well guys, Those who visited HAR2009, especially who visited the GSM related events, how was it? Share us you adventures :)

2 2

new patch for fragmenting HFC-E1
by Andreas.Eversberg 24 Aug '09

24 Aug '09

hi, here is the almost final patch for fragmenting HFC-E1 cards for multiple BTS connecting to one card. if it is stable, i will submit it after some tests. load module with: $ modprobe hfcmulti dmask=0x00000042 bmask=0x0000003c,0x00000780 debug=0x40000 the dmask must be given von all E1 cards that have to be fragmented. in this case we only fragment the first card. the bmask must be given for all fragments. because we have two bits set (slot 1 and slot 6), we need two values. each bmask value will define the the bchannels to use for the given dchannel. the card is fragmented from lower dchannel to higher dchannel given in the mask. (slot 1 for first fragment, slot 2 for second fragment) run misdn_info after loading to see the fragments. happy testing, andreas

2 1

some bug reports for subscr usage
by Andreas.Eversberg 23 Aug '09

23 Aug '09

hi, i found some leaking on the subscriber ressource usage. this patch is not final. maybe we find better solutions. 1. trans_alloc() may not allocate subscriber for transaction, because may be already done before, so i removed subscr_get() from transaction.c -> on outgoing calls we allocate subscriber ressource (by searching for it) and use it for transaction. -> in incomming calls we use the subscriber pointer from the lchan. 2. on an incomming call/sms, the lchan holds a subscriber ressource, so i added subscr_get() after successfully calling trans_alloc() 3. when _send_sms_buffer() is called, we already hold a subscriber ressource for the receiver, so the subsequently called function sms_from_text() must not allocate the receiver again. if this function fails, the allocated subcriber ressource must be released in _send_sms_buffer(). 4. on trans_free we don't need to remove pending paging processes, because it is done when paging expires. anyway if we want to keep it in trans_free(), we must remove paging_remove_request() from paging_T3113_expired(), because it will then be done when transaction is removed. also we need to add paging_remove_request() when sms is removed due to paging expiery. i also fixed linux call router for compiling with openbsc without any patch at all. by linking the source directory to the lcr source directory, the "configure" script automatically detects it and it will compile. (latest git) regards, andreas

1 0

Does anyone have IuH (RUA/HNBAP) captures?
by Harald Welte 23 Aug '09

23 Aug '09

Hi! I've alerady sent this to wireshark-dev, but maybe somebody on this list can also help... I'm right now working for a dissector for the 3GPP IuH (Home-NodeB) protocol. However, I believe the pcap files that I currently have for the protocol are actually invalid. Is there anyone on this list who would be able to provide me with RUA and/or HNBAP protocol traces (preferrably inside SCTP)? Thanks a lot. -- - 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

BS-11 link does not come up
by Sergey V. Efimoff 22 Aug '09

22 Aug '09

Hello All, When trying to run bsc_hack I get the following messages: bootstrapping OML for BTS 0 <0020> abis_nm.c:1644 Set BTS Attr (bts=0) <0020> abis_nm.c:1661 Set TRX Attr (bts=0,trx=0) <0020> abis_nm.c:1754 Set Chan Attr (bts=0,trx=0,ts=0) <0020> abis_nm.c:1754 Set Chan Attr (bts=0,trx=0,ts=1) <0020> abis_nm.c:1624 CONNECT TERR TRAF Um=(bts=0,trx=0,ts=1) E1=(0,2,1) <0020> abis_nm.c:1754 Set Chan Attr (bts=0,trx=0,ts=2) <0020> abis_nm.c:1624 CONNECT TERR TRAF Um=(bts=0,trx=0,ts=2) E1=(0,2,2) <0020> abis_nm.c:1754 Set Chan Attr (bts=0,trx=0,ts=3) <0020> abis_nm.c:1624 CONNECT TERR TRAF Um=(bts=0,trx=0,ts=3) E1=(0,2,3) <0020> abis_nm.c:1754 Set Chan Attr (bts=0,trx=0,ts=4) <0020> abis_nm.c:1624 CONNECT TERR TRAF Um=(bts=0,trx=0,ts=4) E1=(0,3,0) <0020> abis_nm.c:1754 Set Chan Attr (bts=0,trx=0,ts=5) <0020> abis_nm.c:1624 CONNECT TERR TRAF Um=(bts=0,trx=0,ts=5) E1=(0,3,1) <0020> abis_nm.c:1754 Set Chan Attr (bts=0,trx=0,ts=6) <0020> abis_nm.c:1624 CONNECT TERR TRAF Um=(bts=0,trx=0,ts=6) E1=(0,3,2) <0020> abis_nm.c:1754 Set Chan Attr (bts=0,trx=0,ts=7) <0020> abis_nm.c:1624 CONNECT TERR TRAF Um=(bts=0,trx=0,ts=7) E1=(0,3,3) bootstrapping RSL for BTS/TRX (0/0) using MCC=1 MNC=1 BSIC=1 TSC=7 <0020> abis_nm.c:805 OC=SIEMENSHW(a5) INST=(03,00,00) STATE CHG: <0020> abis_nm.c:805 OC=SIEMENSHW(a5) INST=(03,00,00) Failure Event Report Severity=indeterminate failure <0020> abis_nm.c:805 OC=SIEMENSHW(a5) INST=(03,00,00) STATE CHG: Debug with -dDMI:DRSL:DNM shows strange lines: ... <1000> input/misdn.c:132 DL_INFORMATION_IND: use channel(0) sapi(63) tei(127) for now mISDN message for unknown sign_link ... <1000> input/misdn.c:132 DL_INFORMATION_IND: use channel(2) sapi(62) tei(1) for now mISDN message for unknown sign_link ... Why TEI=1 or TEI=127 while it should be 25? Also, dmesg produces: ... mISDN_send: error -12 ... I've reviewed mailing list for similiar problems description and found that several people have already seen these problems, but didn't find the solution. I thought that these problems may be connected with internal BS11 oscillator calibration, so I've performed calibration as described in wiki (though I used Nateks SDH mux with Stratum oscillator as E1 clock source instead of HFC-S ISDN card, which I don't have). Calibration was successful, but it didn't help. Here is query output while running bsc_hack: ... LMT LOGON: ACK PHASE: 3 Normal MBCCU0: Load MBCCU1: No Load Abis-link: Up BS11 ATTRIBUTES: BS-11 ESN PCB Serial Number: 001132 BS-11 ESN Hardware Code Number: 135-2044/03.03 BS-11 ESN Firmware Code Number: 135-2044/03.03 E1 Channel: Port=255 Timeslot=161 Subslot=39 PLL Set Value=1024, Work Value=1023 SITE MANAGER ATTRIBUTES: E1 Channel: Port=0 Timeslot=1 (Full Slot) TEI: 25 BS11 Line Interface ATTRIBUTES: PLL Mode: Standalone BS11 CCLK ATTRIBUTES: CCLK Accuracy: Medium (0) BS11 Power Amplifier 0 ATTRIBUTES: TRX Power: 30mW (GSM) LMT LOGOFF: ACK ... It seems that everything should be ok, but... there is no success. Best regards, Sergey.

2 4

RFC: Change gsm_subscriber tmsi from char[17] to u_int32_t
by Holger Freyther 22 Aug '09

22 Aug '09

Hey guys, for some historic reason our tmsi is an character array but all places dealing with it treat it as a 4 octet, and according to the spec it is a four octet. I have changed the definition and the code using the tmsi. It can be found in my onwaves.com branch. Does it make sense to switch? z.

3 2

Jump to page:

2026

2025

2024

2023

2022

2021

2020

2019

2018

2017

2016

2015

2014

2013

2012

2011

2010

2009

OpenBSC