This issue should already be fixed in current darcs. In FreeSWITCH our copy of sofia is currently sync'd with darcs and it is more stable than 1.12.9 for sure and likely 1.12.8 as well.
Mike Mike On Nov 25, 2008, at 11:58 AM, Jen Chitty wrote: > Hi, > > I'm running 1.12.9 and I'm finding that my process leaks memory > when it is the caller (sends the INVITE) and it terminates the > call (sends the BYE). However, if the callee terminates the > call, there's no leak. > > This is a blocker bug for us, and I don't have a workaround. > I would really appreciate any assistance that anyone has to offer. > > Here's the SOFIA_DEBUG 9 trace from a call where the caller sends > the BYE (NOTE: the log was collected on the caller side): > > nua: nh_create_handle: entering > nua: nua_handle_bind: entering > nua: nua_invite: entering > nua: nua_stack_set_params: entering > soa_clone(static::0x1001a120, 0x100177f8, 0x10024280) called > soa_set_params(static::0x10025268, ...) called > soa_set_user_sdp(static::0x10025268, (nil), 0x10018eff, -1) called > soa_set_capability_sdp(static::0x10025268, (nil), 0x10018eff, -1) > called > su_localinfo: if lo with index 1 > su_localinfo: if bridget with index 4 > soa_set_params(static::0x10025268, ...) called > nta_leg_tcreate(0x100256d8) > nua(0x10024280): adding session usage > soa_init_offer_answer(static::0x10025268) called > soa_generate_offer(static::0x10025268, 0) called > soa_static_offer_answer_action(0x10025268, soa_generate_offer): called > soa_static(0x10025268, soa_generate_offer): generating local > description > su_localinfo: if lo with index 1 > su_localinfo: if bridget with index 4 > soa_static(0x10025268, soa_generate_offer): upgrade with local > description > soa_sdp_mode_set(0x7f5fd868, (nil), ""): called > soa_static(0x10025268, soa_generate_offer): storing local description > soa_get_local_sdp(static::0x10025268, [(nil)], [0x7f5ff990], > [0x7f5ff994]) > called > nta: selecting scheme sip > tport_tsend(0x10022318) tpn = */192.168.0.152:5060 > tport_resolve addrinfo = 192.168.0.152:5060 > tport_by_addrinfo(0x10022318): not found by name */192.168.0.152:5060 > tport_vsend returned 685 > nta: sent INVITE (10828982) to */192.168.0.152:5060 > tport_pend(0x10022318): pending 0x100264f0 for udp/192.168.0.151:5060 > (already 0) > nta: timer set to 32000 ms > nta: timer shortened to 500 ms > nua(0x10024280): call state changed: init -> calling, sent offer > soa_get_local_sdp(static::0x10025268, [0x7f5ff9a8], [0x7f5ff9ac], > [(nil)]) > called > nua: nua_application_event: entering > nua(0x10024280): sent signal r_invite > tport_wakeup_pri(0x10022318): events IN > tport_recv_event(0x10022318) > tport_recv_iovec(0x10022318) msg 0x10027d00 from (udp/ > 192.168.0.151:5060) > has 281 bytes, veclen = 1 > tport_deliver(0x10022318): msg 0x10027d00 (281 bytes) from > udp/192.168.0.152:5060/sip next=(nil) > nta: received 100 Trying for INVITE (10828982) > nta: 100 Trying is going to a transaction > nta_outgoing: RTT is 20 ms > tport_release(0x10022318): 0x100264f0 by 0x10027328 with 0x10027d00 > (preliminary) > tport_wakeup_pri(0x10022318): events IN > tport_recv_event(0x10022318) > tport_recv_iovec(0x10022318) msg 0x10027d00 from (udp/ > 192.168.0.151:5060) > has 486 bytes, veclen = 1 > tport_deliver(0x10022318): msg 0x10027d00 (486 bytes) from > udp/192.168.0.152:5060/sip next=(nil) > nta: received 180 Ringing for INVITE (10828982) > nta: 180 Ringing is going to a transaction > tport_release(0x10022318): 0x100264f0 by 0x10027328 with 0x10027d00 > (preliminary) > nua: nua_application_event: entering > nua(0x10024280): call state changed: calling -> proceeding > nua: nua_application_event: entering > nta: timer not set > tport_wakeup_pri(0x10022318): events IN > tport_recv_event(0x10022318) > tport_recv_iovec(0x10022318) msg 0x10028510 from (udp/ > 192.168.0.151:5060) > has 664 bytes, veclen = 1 > tport_deliver(0x10022318): msg 0x10028510 (664 bytes) from > udp/192.168.0.152:5060/sip next=(nil) > nta: received 200 OK for INVITE (10828982) > nta: 200 OK is going to a transaction > tport_release(0x10022318): 0x100264f0 by 0x10027328 with 0x10028510 > nta: timer set to 32000 ms > soa_set_remote_sdp(static::0x10025268, (nil), 0x10028a3c, 132) called > soa_process_answer(static::0x10025268) called > soa_static_offer_answer_action(0x10025268, soa_process_answer): called > soa_sdp_mode_set(0x10026cf8, 0x10028148, ""): called > soa_static(0x10025268, soa_process_answer): upgrade codecs with remote > description > soa_activate(static::0x10025268, (nil)) called > nua(0x10024280): INVITE: processed SDP answer in 200 OK > nua: nua_application_event: entering > soa_activate(static::0x10025268, (nil)) called > nta: selecting scheme sip > tport_tsend(0x10022318) tpn = */192.168.0.152:5060 > tport_resolve addrinfo = 192.168.0.152:5060 > tport_by_addrinfo(0x10022318): not found by name */192.168.0.152:5060 > tport_vsend returned 306 > nta: sent ACK (10828982) to */192.168.0.152:5060 > nua(0x10024280): call state changed: proceeding -> ready, received > answer > soa_get_remote_sdp(static::0x10025268, [0x7f5ff6b0], [0x7f5ff6b4], > [(nil)]) called > soa_get_params(static::0x10025268, ...) called > nua: nua_application_event: entering > nua: nua_application_event: entering > nua: nua_handle_bind: entering > nua: nua_bye: entering > nua(0x10024280): sent signal r_bye > nua: nua_stack_set_params: entering > soa_set_params(static::0x10025268, ...) called > soa_terminate(static::0x10025268) called > soa_init_offer_answer(static::0x10025268) called > nta: selecting scheme sip > tport_tsend(0x10022318) tpn = */192.168.0.152:5060 > tport_resolve addrinfo = 192.168.0.152:5060 > tport_by_addrinfo(0x10022318): not found by name */192.168.0.152:5060 > tport_vsend returned 451 > nta: sent BYE (10828983) to */192.168.0.152:5060 > tport_pend(0x10022318): pending 0x10025918 for udp/192.168.0.151:5060 > (already 0) > nta: timer shortened to 500 ms > tport_wakeup_pri(0x10022318): events IN > tport_recv_event(0x10022318) > tport_recv_iovec(0x10022318) msg 0x1002a970 from (udp/ > 192.168.0.151:5060) > has 416 bytes, veclen = 1 > tport_deliver(0x10022318): msg 0x1002a970 (416 bytes) from > udp/192.168.0.152:5060/sip next=(nil) > nta: received 200 OK for BYE (10828983) > nta: 200 OK is going to a transaction > nta_outgoing: RTT is 70 ms > tport_release(0x10022318): 0x10025918 by 0x1002a3a0 with 0x1002a970 > nua: nua_application_event: entering > nua(0x10024280): call state changed: terminating -> terminated > nua: nua_application_event: entering > nua: nua_application_event: entering > nua: nua_handle_destroy: entering > nua(0x10024280): sent signal r_destroy > nua(0x10024280): removing session usage > soa_destroy(static::0x10025268) called > nta_leg_destroy(0x100256d8) > nua: terminated session 0x10024280 > nta_leg_destroy((nil)) > nta: timer set next to 4540 ms > nta: timer K fired, terminate BYE (10828983) > outgoing_reclaim_all((nil), (nil), 0x7f5ffb78) > nta_outgoing_timer: 0/0 resent, 0/1 tout, 1/2 term, 1/3 free > nta: timer set next to 22520 ms > nta: timer D fired, terminate INVITE (10828982) > nta: timer F fired, terminating ACK (10828982) > outgoing_reclaim_all((nil), (nil), 0x7f5ffb78) > nta_outgoing_timer: 0/0 resent, 1/1 tout, 1/1 term, 2/2 free > nta: timer not set > > I looked into the home-based allocation code and came across > the statistics code. I hacked in some code to enable the > statistics on the nua_t object's home just after nua_create > returns. If I check the memory statistics on the nua_t > object's memory home after carrying out a call in which > the caller sends the BYE (including waiting for all timers > to expire), I see the following on the caller side: > > hs_clones: 1 > hs_rehash: 0 > hs_blocksize: 31 > hsp_size: 0 > hsp_used: 0 > hsa_number: 1 > hsa_bytes: 148 > hsa_rbytes: 152 > hsa_maxrbytes: 152 > hsa_preload: 0 > hsf_number: 0 > hsf_bytes: 0 > hsf_rbytes: 0 > hsf_preload: 0 > hsb_number: 1 > hsb_bytes: 148 > hsb_rbytes: 152 > > But, if I restart and carry out a call in which the callee > sends the BYE, then I see the following instead (on the > caller side, after all timers expire): > > hs_clones: 1 > hs_rehash: 0 > hs_blocksize: 31 > hsp_size: 0 > hsp_used: 0 > hsa_number: 1 > hsa_bytes: 148 > hsa_rbytes: 152 > hsa_maxrbytes: 152 > hsa_preload: 0 > hsf_number: 1 > hsf_bytes: 148 > hsf_rbytes: 152 > hsf_preload: 0 > hsb_number: 0 > hsb_bytes: 0 > hsb_rbytes: 0 > > It certainly looks to me like something isn't getting freed > in the case where the caller sends the BYE. > > Should I go back to 1.12.8? I read that someone else had gone > back to 1.12.8 to avoid memory leaks in 1.12.9. I could try > grabbing the latest from Darcs, but I'm nervous about using > that, as it's not a "stable release", and I really do need > stability. > > Thanks. > > --Jen > > Jen T'ai Chitty > Chief Software Engineer, VoIP > VTech Technologies Canada, Ltd. > +1-604-273-5131 ext 849 > > ------------------------------------------------------------------------- > This SF.Net email is sponsored by the Moblin Your Move Developer's > challenge > Build the coolest Linux based applications with Moblin SDK & win > great prizes > Grand prize is a trip for two to an Open Source event anywhere in > the world > http://moblin-contest.org/redirect.php?banner_id=100&url=/ > _______________________________________________ > Sofia-sip-devel mailing list > Sofia-sip-devel@lists.sourceforge.net > https://lists.sourceforge.net/lists/listinfo/sofia-sip-devel ------------------------------------------------------------------------- This SF.Net email is sponsored by the Moblin Your Move Developer's challenge Build the coolest Linux based applications with Moblin SDK & win great prizes Grand prize is a trip for two to an Open Source event anywhere in the world http://moblin-contest.org/redirect.php?banner_id=100&url=/ _______________________________________________ Sofia-sip-devel mailing list Sofia-sip-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/sofia-sip-devel
