- OpenBSC - lists.osmocom.org

ChanModify
by Pierre Kim 11 Nov '12

11 Nov '12

<000f> bsc_api.c:315 Sending ChanModify for speech 1 1 <000c> rtp_proxy.c:258 Correcting frame difference of 38886 frames I used to get these messages when I make a call using osmo-nitb. But recently I updated to newest git version and I get something like <000b> bsc_api.c:374 Sending ChanModify for speech 0 1 <000b> bsc_api.c:118 Using non speech mode: 0 <000b> osmo_msc.c:73 Assignment complete should not have been reached. And I get no sound for my phones. There is a year gap between the previous version and the new version, so I couldn't locate exactly what change is causing this. Can somebody help me fixing it?

3 24

29C3 osmo-nitb sprint and issues
by Holger Hans Peter Freyther 08 Nov '12

08 Nov '12

Hi, as indicated in the previous mail there are some things to be done before we get to the venue. I would like to experiment with sprints to solve some of the stability issues we had with osmo-nitb. The idea is that we assemble (physically or online) and work on some specific issues of osmo-nitb that impact the quality of service for the upcoming event. I can provide food and snacks during these sprints and guidance on OpenBSC code. The first sprint is scheduled for the 9th, 10th (and maybe 11th) of November. There is no venue for it yet, maybe the CCCB. Please indicate if you are interested to join. The general list of osmo-nitb issues are: * Channel release process. There is a branch I created at the last congress but I have never tested it. This includes making a SAPI=3 local release, using the proper timeouts for the abnormal release making sure the SACCH is deactivated and some more issues. This requires quite a bit of testing. * Allow to block/unblock failing lchans. We have had this issue with Nokia BTS but it can happen on other BTS as well. E.g. a given timeslot might not be allocatable, or we get LCHAN ACT NACK. In that case OpenBSC should block the lchan. * smsqueue/auth/paths. In the past years we have seen that the smsqueue and other parts can get stuck because of the 'subscr_get_con' not being properly called in all error conditions. We need to review the code and check how things get stuck. * Handover testing and patching out the 'timestamp' adjusting or fixing it. Right now it can be used to add a lot of delay by switching cells. cheers holger

1 2

Re[2]: Non-synchronized handover
by Dmitri Soloviev 05 Nov '12

05 Nov '12

as for OpenBTS handover, I studied OpenBSC code. I'm issuing exactly the same handover command, even including the sequence of fields, not only values :) I'm waiting for the second umTXR back to continue investigating.. Mon, 5 Nov 2012 13:34:33 -0500 от Alexander Chemeris <alexander.chemeris(a)gmail.com>: > > > > >Hi Harald, > > On Thu, Nov 1, 2012 at 2:22 PM, Harald Welte <laforge(a)gnumonks.org> wrote: > > On Thu, Nov 01, 2012 at 11:33:28AM +0400, Alexander Chemeris wrote: > >> Dimitri, we have a working OpenBSC/nanoBTS installation at the hackerspace. > >> If you find out how to issue a handover command in OpenBSC, you could try > >> by yourself. > > > > You cannot issue a handover command manually in OpenBSC. The way how we > > test handover is to use two BTSs connected to one OsmoNITB, and use > > rotary/adjustable attenuators in the Tx path to make one BTS signal > > weaker than the other BTS, at which point the handover algorithm will > > kick in and perform the hand-over. > > If this is done inside of the OpenBSC, it should be easy to hack a VTY > command to force handover. It may be a useful command as well, if you > want to do load balancing of your network. Am I missing something? > > > The DMEAS and the DHO debug logging might be useful to monitor what is > > happening while you're doing that. > > Thanks. > > -- > Regards, > Alexander Chemeris. > CEO, Fairwaves LLC / ООО УмРадио >http://fairwaves.ru >

1 0

Non-synchronized handover
by Dmitri Soloviev 05 Nov '12

05 Nov '12

Hi, I'm implementing handover for OpenBTS project: http://wush.net/trac/rangepublic/wiki/Handover Some handsets perform handover properly, while the others (such as Nokia) do not appear at the target channel with HandoverAccess. Handover Command contains the minimal set of parameters that looks reasonable: -- cellDescription -- channelDescription -- handoverReference -- powerCommand - tried various levels I'll appreciate real-life traces of handover, non-sync case. As far as I know, this one is used for ipAccess.. Handover Command fields are particularly interesting. Thank you, Dmitri

3 5

Bounce action notification
by mailman＠lists.osmocom.org 05 Nov '12

05 Nov '12

This is a Mailman mailing list bounce action notice: List: OpenBSC Member: openbsc(a)slewe.com Action: Subscription disabled. Reason: Excessive or fatal bounces. The triggering bounce notice is attached below. Questions? Contact the Mailman site administrator at mailman(a)lists.osmocom.org.

1 0

triplets calculation
by Yann R. Moupinda 01 Nov '12

01 Nov '12

Hi guys, i'm trying to interconnect a WLAN router with my GSM-sysmoBTS and to use the EAP-SIM authentication protocol. For that i installed a Freeradius on an external board that i use as radius server. The Freeradius should use a database file in order to realise the authentication and the idea is that Freeradius should be able to read the hlr.sqlite3 file from my sysmoBTS. I'm not so far to do this part of the work so, i first want to do some tests using the sim card from sysmoBTS (imsi und ki) and the triplets (RAND, Kc and SRES). I want to create a flat file containing this values so the Freeradius will directly use the file to get the triplets. I thought i could get the triplets stored in one of the tables of hlr.sqlite3 (AuthLastTuples) but there is no information on it. From the table AuthKeys i could find some Ki values (e.g X'019B7083FBAFC928421A147DE795217782'). I found that there is an "osmo-auc-gen" program for computing the triplets. By using the Ki from AuthKeys and a self generated RAND (e.g 0123456789ABCDEF0123456789ABCDEF) the program doesn't run and doesn't give the triplets. The used authentication algorithm is the COMP128v1. Has anyone an idea why it doesn't work and why there is no information in the AuthLastTuples table from sysmoBTS ? best regards Yann

2 2

[openbsc PATCH v2] Make openbsc also compile/run on OSX
by Tobias Engel 30 Oct '12

30 Oct '12

I modified the patch for ipaccess-find based on Jim's and Holger's suggestions. Still not sure what to do about the (coming) librt dependency, so no new version of that patch for now. Tobias Tobias Engel (1): Remove iface arg if SO_BINDTODEVICE isn't available openbsc/src/ipaccess/ipaccess-find.c | 13 +++++++++++-- 1 file changed, 11 insertions(+), 2 deletions(-) -- 1.7.10.2 (Apple Git-33)

1 1

BS-11 and HFC-E1 for sale
by Bernd Baumgraß 30 Oct '12

30 Oct '12

Hello, I am thinking about selling my Siemens BS-11 and a compatible HFC-E1 Evaluation Board. If anyone is interested please send me a mail: baumgrass at hotmail.com . Regards, Bernd.

1 0

Testing OpenBSC with a FakeBTS
by Holger Hans Peter Freyther 28 Oct '12

28 Oct '12

Hi, I wrote some code that lets me test OpenBSC without using real hardware, e.g. one is bothered to wait for the nanoBTS to open the OML connection again, one doesn't have the unit for the right band, one doesn't have a dual trx system, or because reproducing the issue is difficult (RF interference, LAPDm timeouts). I have packaged the code with OBS and right now it is only available for debian stable (it can be installed on unstable as long as the stable feeds are inside the sources.list). Install: $ echo 'deb http://download.opensuse.org/repositories/home:/zecke23/Debian_6.0/ ./' > /etc/apt/sources.list.d/ $ aptitude update; aptitude install gnu-smalltalk osmo-st-openbsc-test (the packages are signed but the key expired two years ago, so the question is if you want to trust Debian packages from OpenSUSE) Using: $ gst st> PackageLoader fileInPackage: #FakeBTS ... st> bts := FakeBTS.BTS new. st> bts btsId: '1/0/0' st> bts connect: 'localhost'; waitForBTSReady. ... Testing: There is the FakeBTS.OpenBSCTest class that has methods to require channels and these Lchans can be used to send RSL messages. I don't have a public example at this point though. Vision: I hope this code will be used to do nightly tests of OpenBSC, specially the error paths (e.g. timeouts on handover, timeouts on assignment, timeouts on channel release, etc.) There is a branch that brings DualTRX support and I am going to merge this to master soon. What I am missing is some test launcher code (it can be in python or ruby if someone volunteers) that will launch osmo-bsc/osmo-nitb with a given configuration file and initialize the HLR with a set of given subscribers. enjoy holger PS: Blog post will follow once the code has matured a bit

1 1

[openbsc PATCH 0/3] Make openbsc also compile/run on OSX
by Tobias Engel 27 Oct '12

27 Oct '12

These are some minor modifications to make openbsc run on OSX. You have to have these MacPorts installed: pkgconfig automake autoconf libtool libdbi libdbi-drivers sqlite3 ... and set: export LDFLAGS='-L/opt/local/lib' export CPPFLAGS='-I/opt/local/include' Everything else should work as usual. Tobias Engel (3): Remove interface argument when compiled on OSX Remove unused librt dependency Set byte order defines when compiled on OSX openbsc/src/ipaccess/ipaccess-find.c | 4 ++++ openbsc/src/libmgcp/mgcp_network.c | 8 +++++++- openbsc/src/libtrau/rtp_proxy.c | 8 +++++++- openbsc/tests/mgcp/Makefile.am | 2 +- 4 files changed, 19 insertions(+), 3 deletions(-) -- 1.7.10.2 (Apple Git-33)

5 11