Hi Peter,

Ok. Do you have a later message from the following line?

LOGP(DLLAPD, LOGL_INFO, "I received in state %s\n", lapd_state_names[dl->state]);

Sure, this is how it looks with all relevant debugging enabled:

Fri Aug 2 16:50:13 2013 <0018> lapd_core.c:1465 I received in state LAPD_STATE_MF_EST Fri Aug 2 16:50:13 2013 <0018> lapd_core.c:1548 incrementing V(R) to 1 Fri Aug 2 16:50:13 2013 <0018> lapd_core.c:222 stop T203 Fri Aug 2 16:50:13 2013 <0018> lapd_core.c:206 start T203 Fri Aug 2 16:50:13 2013 <0018> lapd_core.c:1557 message in single I frame Fri Aug 2 16:50:13 2013 <001b> e1_input.c:594 RX: 80 80 00 08 81 00 00 01 5f 23 01 01 sapi=62 tei=1 Fri Aug 2 16:50:13 2013 <0005> bts_nokia_site.c:1668 Nokia: Abis received! Fri Aug 2 16:50:13 2013 <0005> bts_nokia_site.c:1688 ABIS_OM_MDISC_FOM Fri Aug 2 16:50:13 2013 <0005> bts_nokia_site.c:1515 (0x81) NOKIA_BTS_ACK Fri Aug 2 16:50:13 2013 <0005> bts_nokia_site.c:1547 ACK = 1 Fri Aug 2 16:50:13 2013 <0005> bts_nokia_site.c:1588 Nokia: STOP OML & RSL lapd (cause: RESET) Fri Aug 2 16:50:13 2013 <0018> input/lapd.c:513 LAPD DL-RELEASE request TEI=1 SAPI=62 Fri Aug 2 16:50:13 2013 <0018> lapd_core.c:2183 Message DL-RELEASE-REQUEST received in state LAPD_STATE_MF_EST Fri Aug 2 16:50:13 2013 <0018> lapd_core.c:2024 perform local release Fri Aug 2 16:50:13 2013 <0018> lapd_core.c:229 new state LAPD_STATE_MF_EST -> LAPD_STATE_IDLE Fri Aug 2 16:50:13 2013 <0018> lapd_core.c:222 stop T203 Fri Aug 2 16:50:13 2013 <0018> input/lapd.c:628 LAPD DL-RELEASE confirm TEI=1 SAPI=62 Fri Aug 2 16:50:13 2013 <001b> e1_input.c:585 DL_REL: sapi(62) tei(1) Fri Aug 2 16:50:13 2013 <001b> bsc_init.c:318 Lost some E1 TEI link: 1 0x887f038 Fri Aug 2 16:50:13 2013 <0018> lapd_core.c:1781 lapd_send_i() called from line 1615 Segmentation fault (core dumped)

bsc_init.c probably complaining because the LAPD connection is closed, but the OML context does not. Maybe this is the root cause of the problem: if the OML context not disconnected with the LAPD connection, it still listens on the OML timeslot and if something got received, it tries to process it. But still: this is not how it should work. The OML context is not closed, because we are waiting for the already resetting BTS, then the OML lapd got restarted again (after the RESET_TIMER expired). My opinion is that after we closed the lapd connection, it should not reactivate itself, if data appears on the OML TS.

BR, Csaba