Hi,

I already mentioned a strange behavior, sometimes when running osmocom for a few days. I think this is a different issue, because it happens more frequently and the situation is a bit different. This problem does not occur with a binary from june 2012. So it has to be some change between june and now.

But now the problem: When running a phone with the main branch (only tested), it sometime come to the situation that the phone says "connection pending" and is not reacting anymore. osmocon only logs periodically this line:

TOA AVG is not 16 qbits, correcting (got 15)

And mobile says:

<0003> gsm322.c:474 Sync to ARFCN=694(DCS) rxlev=-80 (No sysinfo yet, ccch mode NONE) <000e> gsm48_mm.c:353 Periodic location update <0005> gsm48_mm.c:355 timer T3212 (periodic loc. upd. delay) has fired <0005> gsm48_mm.c:4338 (ms 1) Received 'MM_EVENT_TIMEOUT_T3212' event in state MM IDLE, normal service <000e> gsm48_mm.c:2222 Perform location update (MCC 262, MNC 07 LAC 0x27e9) <0005> gsm48_mm.c:2356 LOCATION UPDATING REQUEST <0005> gsm48_mm.c:2378 using LAI (mcc 262 mnc 07 lac 0x27e9) <0005> gsm48_mm.c:2386 using TMSI 0xXXXXXXXX <0005> gsm48_mm.c:917 new state MM IDLE, normal service -> wait for RR connection (location updating) <0001> gsm48_rr.c:5575 (ms 1) Message 'RR_EST_REQ' received in state idle (sapi 0) <000e> gsm48_rr.c:1352 Establish radio link due to mobility management request <0003> gsm322.c:4049 (ms 1) Event 'EVENT_LEAVE_IDLE' for Cell selection in state 'C3 camped normally' <0003> gsm322.c:829 new state 'C3 camped normally' -> 'connected mode 1' <0003> gsm322.c:3665 Going to camping (normal) ARFCN 664(DCS). <0003> gsm322.c:452 Sync to ARFCN=664(DCS), but there is a sync already pending <0001> gsm48_rr.c:355 new state idle -> connection pending <0001> gsm48_rr.c:1504 CHANNEL REQUEST: 00 (Location Update no NECI)

Mobile says nothing when shutting the phone off or on:

OsmocomBB(ms)#shutdown OsmocomBB(ms)# OsmocomBB(ms)#no shutdown OsmocomBB(ms)#

Only logs:

<0005> gsm48_mm.c:4342 (ms 1) Received 'MM_EVENT_IMSI_DETACH' event in state wait for RR connection (location updating) <0005> gsm48_mm.c:1992 IMSI detach delayed.

Killing and restarting mobile leads to:

eeepc:~ # mobile -i 127.0.0.1 Copyright (C) 2008-2010 ... Contributions by ...

License GPLv2+: GNU GPL version 2 or later http://gnu.org/licenses/gpl.html This is free software: you are free to change and redistribute it. There is NO WARRANTY, to the extent permitted by law.

<000f> sim.c:1223 init SIM client <0006> gsm48_cc.c:63 init Call Control <0007> gsm480_ss.c:231 init SS <0017> gsm411_sms.c:63 init SMS <0001> gsm48_rr.c:5626 init Radio Ressource process <0005> gsm48_mm.c:1327 init Mobility Management process <0005> gsm48_mm.c:1040 Selecting PLMN SEARCH state, because no SIM. <0002> gsm322.c:5037 init PLMN process <0003> gsm322.c:5038 init Cell Selection process <0003> gsm322.c:5095 Read stored BA list (mcc=262 mnc=01 Germany, T-Mobile) <0003> gsm322.c:5095 Read stored BA list (mcc=262 mnc=07 Germany, O2) Mobile '1' initialized, please start phone now! VTY available on port 4247.

And layer1 still logs some:

TOA AVG is not 16 qbits, correcting (got 15) TOA AVG is not 16 qbits, correcting (got 15) TOA AVG is not 16 qbits, correcting (got 15)

Has anybody a clue how this could happen from time to time?

Cheers

Tim