Hi zecke and others,

I may have found a cause of one of the problems that we've been experiencing in nanoBTS bring-up, at least at some point in the past.

The problem could be observed/described as: During the first OML connect of the nanoBTS, OpenBSC fails to bring it up completely. Second and successive connects worked fine.

What I have just noticed is that we never re-set the NM state information for the 12.21 MOs inside OpenBSC. This means, if the A-bis link is lost, the old state information continues to exist.

I guess it would be cleaner to simply re-set all the nm_state strucures that are part of the 'struct gsm_bts', gsm_bts_trx, etc. when the A-bis link is lost.

As I'm working in this area right now anyway, I intend to produce a patch addressing the issue.

But now that I think more about it: I guess it is unrelated to the old problem, as a typical work-around was to re-start OpenBSC, at which time all state is re-set again.

Regards, Harald