Hi!
I've now merged the 'laforge/hopping' branch to master, and wanted to update you on the state of frequency hopping support:
First, the bad news: It is still not working with the BS-11 :(
All OML attributes that I can think of are set correctly, they have been verified from wireshark, and the BTS acknowledges all those attributes.
On RSL, the channels are activated the right way, and even the SYSTEM INFORMATIO 1 (cell channel allocation) as well as the chennel description and mobile allocation parts of the IMMEDIATE ASSIGN are encoded correctly.
Still, the MS and BTS fail to establish any hopping channel. Dieter is now trying to get the BS-11 hopping configuration working with his Racal 6113. If that works, the protocol trace should reveal where we still do something wrong.
If you want to play with it: The BTS attributes as well as the TRX attributes are not yet 100% finished like they should be. So you need to apply the attached patch as a hack on top of master. Please note that the bs11_attr_radio_trx1 contains hard-coded ARFCN 117 and 119, i.e. you will have to modify this unless your hopping sequence also only consists of those two ARFCN.
I've also attached a config file for your reference.
Hi all!
As it turns out, the BTS side of the hopping was working all the time. However, the phones discarded our IMMediate ASSignment message due to an erroneous L2 Pseudo Length.
This has been fixed in git commit cfa4a01c8fe2b5f2ab1bfa6fcf23c35ff5e957f4 and the hopping is working fine using the patch and config that I've sent in the last message.
Please note the BS-11 does synthesizer hopping, so it can hop over any number of channels. I haven't tested more than 8, but you can definitely hop over 8 ARFCN.
Only TRX1 will hop, as TRX0 needs to remain on a constant ARFCN for the CCCH/BCCH. I suggest to use 'channel allocator descending' for testing, as the BSC will then always prefer allocating channels from TRX1, even if suitable channels on TRX0 is still unused.
Thanks to Dieter for verifying the BS-11 actually does hopping using his Racal 6113 analyzer, and thanks to Sylvain for his suggestion to check the L2 pseudo length.
Regards, Harald
-
Harald Welte