Hi Alexander, I fixed the second one :). Before trying the mem debugger, try running 1.6.1 without the stun module - maybe the crashes you see are side effects of the bug in the stun module.
Regards, Bogdan Alexander wrote: > Hello. > > Regarding STUN - it seems that there is the same bug on line 870. > Again, > iterator b2 is created before memory allocation. > > I'll try memory debug, but don't know, when - our OpenSIPS works under > load, > people make calls and so on :) I use version 1.6.0 for now, it seems to be > quite > stable. > > >> Hi Alexander, >> > > >> Indeed, the stun issue looks like a silly bug - I fixed it on SVN - >> Thanks a lot for the report and fix. >> > > >> Regarding the other bugs you reported, could you please compile in the >> memory debug support (http://www.opensips.org/Resources/DocsTsMem) and set >> memlog=8 >> memdump=2 >> > > >> maybe the debugger will be able to get out more info. >> > > >> Regards, >> Bogdan >> > > >> Alexander wrote: >> >>> I've encountered a problem - sometimes OpenSIPS 1.6.0 crashes. GDB >>> says that crash was in stun.c, line 836. Going to this line, we see: >>> >>> case UNKNOWN_ATTRIBUTES: >>> /* iterator */ >>> b2 = (T16 *) msg->unknownAttributes->buffer; >>> /* allocate unknownAttributes buffer */ >>> msg->unknownAttributes = (Buffer*) >>> pkg_malloc(sizeof(Buffer)); >>> if(!msg->unknownAttributes){ >>> LM_DBG("out of mem\n"); >>> return -1; >>> } >>> memset(msg->unknownAttributes, 0, sizeof(Buffer)); >>> >>> Is it OK, that we create iterator, and then allocate buffer? >>> uknownAttributes is not allocated when iterator is created. I've >>> changed it to: >>> >>> case UNKNOWN_ATTRIBUTES: >>> /* allocate unknownAttributes buffer */ >>> msg->unknownAttributes = (Buffer*) >>> pkg_malloc(sizeof(Buffer)); >>> if(!msg->unknownAttributes){ >>> LM_DBG("out of mem\n"); >>> return -1; >>> } >>> memset(msg->unknownAttributes, 0, sizeof(Buffer)); >>> >>> /* iterator */ >>> b2 = (T16 *) msg->unknownAttributes->buffer; >>> >>> Now it seems to work properly. >>> >>> By the way, version 1.6.1 contains the same bug in STUN module. >>> Unfortunately, 1.6.1 is unusable for me - crashes too often :( I wrote >>> one report about crash on NOTIFY (with backtrace), and today I've >>> encountered another crash: >>> >>> Program terminated with signal 11, Segmentation fault. >>> [New process 19330] >>> #0 0x080ec3d9 in fm_malloc (qm=0x81b3c00, size=<value optimized out>) >>> at mem/f_malloc.c:172 >>> 172 *pf=n->u.nxt_free; >>> (gdb) where >>> #0 0x080ec3d9 in fm_malloc (qm=0x81b3c00, size=<value optimized out>) >>> at mem/f_malloc.c:172 >>> #1 0x00603d04 in build_rr (_l=0x81e07d8, _l2=0x81e1d1c, >>> user=0xbff381d0, tag=0x81d1fc0, params=0x0, _inbound=1) at record.c:163 >>> #2 0x0060471f in record_route (_m=0x81d1bec, params=0x0) at record.c:320 >>> #3 0x00606031 in w_record_route (msg=0x81d1bec, key=0x0, bar=0x0) at >>> rr_mod.c:272 >>> #4 0x080545dd in do_action (a=0x81bdb2c, msg=0x81d1bec) at action.c:967 >>> #5 0x08057308 in run_action_list (a=0x81bdb2c, msg=0x81d1bec) at >>> action.c:139 >>> #6 0x080554dd in do_action (a=0x81bdb98, msg=0x81d1bec) at action.c:706 >>> #7 0x08057308 in run_action_list (a=0x81bd578, msg=0x81d1bec) at >>> action.c:139 >>> #8 0x080576a3 in run_top_route (a=0x81bd578, msg=0x81d1bec) at >>> action.c:119 >>> #9 0x0809ddf2 in receive_msg ( >>> buf=0x8192380 "OPTIONS sip:sip.comtube.ru <http://sip.comtube.ru> >>> SIP/2.0\r\nVia: SIP/2.0/UDP >>> 85.21.245.172:5060;branch=z9hG4bK57769584;rport\r\nMax-Forwards: >>> 69\r\nFrom: \"asterisk\" <sip:[email protected] >>> <mailto:sip%[email protected]>>;tag=as58d6d6b9\r\nTo: >>> <sip:sip.comtu"..., len=511, rcv_info=0xbff388d4) at receive.c:162 >>> #10 0x080e5056 in udp_rcv_loop () at udp_server.c:492 >>> #11 0x08070adf in main (argc=5, argv=0xbff38ae4) at main.c:821 >>> ------------------------------------------------------------------------ >>> >>> _______________________________________________ >>> Users mailing list >>> [email protected] >>> http://lists.opensips.org/cgi-bin/mailman/listinfo/users >>> >>> > > > > > > -- Bogdan-Andrei Iancu www.voice-system.ro _______________________________________________ Users mailing list [email protected] http://lists.opensips.org/cgi-bin/mailman/listinfo/users
