| Summary: | ASTERISK-07225: crash on bridging channels, one of which does not have a jitterbuffer enabled | ||
| Reporter: | Sergey Basmanov (sb) | Labels: | |
| Date Opened: | 2006-06-23 08:12:47 | Date Closed: | 2006-06-23 11:52:51 |
| Priority: | Critical | Regression? | No |
| Status: | Closed/Complete | Components: | Core/Jitterbuffer |
| Versions: | Frequency of Occurrence | ||
| Related Issues: | |||
| Environment: | Attachments: | ( 0) ast_bt.txt ( 1) jstorm_bt.5994 | |
| Description: | Also, Zap->H323 (chan_h323) completely crashed PC. I returned back to 1.2.9.1, so only one bt available. Looks like this issue related to jitterbuffer. | ||
| Comments: | By: Jack Storm (jstorm) 2006-06-23 09:53:32 I have the same issue, what follows are the full details: This is /trunk 35630, make valgrind with -DDETECT_DEADLOCKS -DDEBUG_THREADS The segfault for me happend when a callback agent answeres an Agent/ call. jstorm_bt.5994 is my back trace, you will notice that abstract_jb.c line 206 (my bt) and 212 (reporters bt) call ast_test_flag, ast_test_flag ASSUMES that p is initialized. So the question really is, is the problem here in abstract_jb.c or in includes/asterisk/utils.h Also ast_jb_do_usecheck assumes that there is an initialized struct in the pointer for a jitterbuffer in that channel. By: Russell Bryant (russell) 2006-06-23 11:42:01 The changes I did that caused this bug were done in a bad way, so I'm just going to revert them. I'll implement what I intended in a better way next week. By: Russell Bryant (russell) 2006-06-23 11:52:51 breakage reverted, thanks for the report | ||