- OpenBSC - lists.osmocom.org

Re: Calls are "stuck"
by Dieter Spaar 05 Dec '10

05 Dec '10

Hello Andreas, On Sun, 05 Dec 2010 00:28:27 +0100, "Andreas Galauner" <andreas(a)galauner.de> wrote: > > Also, the network sometimes is hard to find on the mobile phones. We > blame a bad calibrated oscillator for that. As soon as we got an HFC-S > card, we will try to recalibrate it. As I mentioned here a few days before, its usually enough to reset the PLL "work value" to its "set value" with bs11_config ("pll_workvalue" command) and set the PLL to standalone ("pll_standalone" command). > Is it possible that all those problems show up because of the > oscillator? And why did it work really well on Thursday? I don't think that the problems with calls are related to the oscillator, I would expect that this is related to the misdn errors you see. Maybe the ISDN cable termination is not set properly ? Best regards, Dieter -- Dieter Spaar, Germany spaar(a)mirider.augusta.de

1 0

Calls are "stuck"
by Andreas Galauner 05 Dec '10

05 Dec '10

Hello everybody, this is my first post to this list. I've been reading it for a while now and I'm finally trying to get a bit deeper into GSM. We have a BS11 in the C4 (Chaos Computer Club Cologne) which I got fully working on Thursday this week. We had several mobile phones in the network and called each other and we could send and receive sms. All fine. Today I wanted to test another few things and everything got weird: I have 2 mobile phones authenticated in the network. One phone can call the other one, it rings and after accepting the call, the first still plays the "ringing sound", while everything on the second is silent. Vice versa, if I call the first mobile from the second, it dials endlessly. Sometimes OpenBSC segfaults. Before it does that, it tells me on the console that timer T303 expired. Another thing are those masses of "mISDN_send: error -12" messages in the kernel log. We got those on Thursday, too but they weren't important to us, because everything worked. Nevertheless, sending and receiving SMS still works. Also, the network sometimes is hard to find on the mobile phones. We blame a bad calibrated oscillator for that. As soon as we got an HFC-S card, we will try to recalibrate it. Is it possible that all those problems show up because of the oscillator? And why did it work really well on Thursday? So far, I tried a few obvious things like rebooting, both the computer and the BS11, cleaning and recompiling everything and I even deleted the old HLR database. I really don't know what to do else. <Insert helpless cry here> ;) Does anybody of you have an idea what I can try? We are running the latest OpenBSC from the git on a Debian testing with a stock 2.6.32-5 kernel on i686. Greetings, Andy

1 0

BS11 power problem
by Sergey V. Efimoff 03 Dec '10

03 Dec '10

Hello All, Herr Thomas had raised up a discussion about problem with setting of BTS power attribute value. In connection with this question, I recall another problem which has occurred with one of my BTS's. It has a very strange behavior. Its output power is set to 2W. My measure device tells me that receive level at the desired channel is about -24 dBm after station boot-up. This value is very good, no other base station around shows better results. But right after the very first GSM call the receive level quickly jumps down to -70 or even -80 dBm, this value corresponds to 30mW of output power. BTS station still works and output power is still shown as 2W. This effect had been discovered with only one of my BTS stations, while other stations work well. Does it mean that it's a hardware problem? Has anyone discovered this problem earlier? Best Regards, Sergey.

2 2

nanoBTS BER Test
by EGU 2100 03 Dec '10

03 Dec '10

Hi, I like to perform a TCH BER Test for nanoBTS via R&S CMU300. Does anyone know how to configure the nanoBTS using ipaccess-config to perform this BER test? Thanks! EGU

2 1

Re: bs11 not working
by Dieter Spaar 02 Dec '10

02 Dec '10

Hello Sergey, On Thu, 2 Dec 2010 21:08:09 +0300, "Sergey V. Efimoff" <risky(a)mail.ru> wrote: > > I had once or twice seen BTS stations which had working PLL value > very different from calibrated value. In order to get them close I had to > use E1 hardware with high precision oscillator, then switch the > hardware to the PLL-locked mode. This operation took me plenty > of time. > > Please tell, is it possible to set the working PLL value directly? Yes, bs11_config supports this. Its the "pll-workvalue" command. The way how the PLL workvalue could be set was not obvious so it took a while till we found out how this could be done. Thats the reason why it took quite long till this command was introduced. Best regards, Dieter -- Dieter Spaar, Germany spaar(a)mirider.augusta.de

1 0

Re: bs11 not working
by Dieter Spaar 02 Dec '10

02 Dec '10

Hello Sylvain, On Thu, 2 Dec 2010 15:20:07 +0100, "Sylvain Munaut" <246tnt(a)gmail.com> wrote: > > The clock of the non working one might just be too off ... > > See http://openbsc.osmocom.org/trac/wiki/isdnsync for a tool to > calibrate (with appropriate HW). > > If you don't have the equipement ... huh, good luck. > You can always try random values until it works. > If you have a USRP you can try to lock onto it and compare offset ... Its possible to set the clock calibration to a fixed value, this was a feature introduced a while ago. The original clock calibration setting workes pretty good, at least from my experience. So if the clock is off due to using the E1 clock as a reference, it should be enough to set the orginal clock calibration value and switch the PLL to standalone mode. Best regards, Dieter -- Dieter Spaar, Germany spaar(a)mirider.augusta.de

2 1

Free Software GSM Talk in Berlin on Dec. 9th
by Ali Gündüz 02 Dec '10

02 Dec '10

Hello all, I'm going to be giving a presentation about free software GSM implementations at the FSFE's Berlin meeting on December 9th, 2010. It will be very elementary and newcomer-friendly. I'm planning to go over OsmocomBB, OpenBTS and OpenBSC, and discuss their security and practical implications. The talk will start at 19:30 in Newthinking Store Tucholskystr. 48, 10117 Berlin. You can find more about it on http://aligunduz.org/blog/free_gsm_talk_next_week.html Considering I'm not an expert on GSM myself, it would be nice to have people more familiar with it in attendance. Ali Gündüz

1 0

Re:socket problem again
by jason mercury 02 Dec '10

02 Dec '10

On 12/01/2010 05:10 PM, Holger Hans Peter Freyther wrote >* we still use the global bsc_gsmnet in many places. We need to remove that to *>* allow something like this.* Is it a too much difficult process? Should i try to do this and where should i start? On 12/01/2010 05:07 PM, Sylvain Munaut wrote: * >That's an option. See nv_put_prim_oml in ipaccess-config.c* I think this is for changing the port of nanoBTS. Where should i modify the port for bsc_hack? Thanks. Jason ** * *

1 0

bs11 not working
by Thomas Ansorg 02 Dec '10

02 Dec '10

hi my 2nd bs11 doesnt work. the boot screen from bs11_config looks like this(attached pic). both boot screens look like that, the operational one too. the 2 bs11 are connected identical, the software is the same. one works, the other doesnt. bad! the bootstrapping works on both, but the handy can only detect the one of two bs11. troubleshooting advices are welcome, openbsc website was no help T. -- Wer Rechtschreibfehler findet, darf sie behalten!

3 2

GPRS Problem
by Stuart Baggs 02 Dec '10

02 Dec '10

Hi, I am openBSC version 0.9.11.92-6d42 and I have compiled open-ggsn and the osmo-sgsn binary has been built when i recompiled openBSC. Now here is the problem. The GPRS indicator on the handset lights up, but no GPRS data passes. I have the following config files: ---------/etc/ggsn.conf--------- # TAG: listen # Specifies the local IP address to listen to listen 192.168.1.70 # TAG: dynip # Dynamic IP address pool. # Used for allocation of dynamic IP address when address is not given # by HLR. # If this option is not given then the net option is used as a substitute. dynip 192.168.254.0/24 # TAG: pcodns1/pcodns2 # Protocol configuration option domain name system server 1 & 2. pcodns1 208.67.222.222 pcodns2 208.67.220.220 -------------------------------- ---------/etc/ggsn.conf--------- ! ! Osmocom SGSN configuration ! ! line vty no login ! sgsn gtp local-ip 192.168.1.80 ggsn 0 remote-ip 192.168.1.70 ggsn 0 gtp-version 1 ! ns timer tns-block 3 timer tns-block-retries 3 timer tns-reset 3 timer tns-reset-retries 3 timer tns-test 30 timer tns-alive 3 timer tns-alive-retries 10 encapsulation udp local-ip 192.168.1.60 encapsulation udp local-port 23000 encapsulation framerelay-gre enabled 0 ! bssgp ! -------------------------------- I have setup the server to have 4 virtual addresses: 192.168.1.50, 192.168.1.60, 192.168.1.70, 192.168.1.80. When I start the ggsn and do a route -n it shows thats tun0 is up with the ip range 192.168.0.0/24. I have specified it to be 192.168.1.254 via the dynip argument in the /etc/ggsn.conf file. I have setup iptables to NAT eth0 so no issues there. Thanks Stuart

2 1