Summary: | ASTERISK-05217: Asterisk sending incorrect facilities to qsig switch in network mode | ||
Reporter: | Paul P Komkoff Jr (stingray) | Labels: | |
Date Opened: | 2005-10-03 02:02:54 | Date Closed: | 2011-06-07 14:10:19 |
Priority: | Minor | Regression? | No |
Status: | Closed/Complete | Components: | Core/General |
Versions: | Frequency of Occurrence | ||
Related Issues: | |||
Environment: | Attachments: | ||
Description: | I have facilityenable = yes switchtype=qsig signalling=pri_net in my zapata.conf. I also have OfficeServ 500 attached. Now, when I'm trying to send something which contain callername facility, the call won't pass. Here's why: > [ 02 01 ea ee 08 02 00 03 05 04 03 80 90 a3 18 03 a1 83 81 1c 1c 9f 8b 01 00 a1 16 02 01 04 02 01 15 30 0e 02 01 01 0a 01 00 a1 02 82 00 a2 02 82 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 6c 0c 00 81 3 8 34 35 32 32 31 30 36 30 32 70 0d 80 39 38 39 31 37 32 30 31 30 34 34 30 ] > Facility (len=30, codeset=0) [ 0x9f, 0x8b, 0x01, 0x00, 0xa1, 0x16, 0x02, 0x01, 0x04, 0x02, 0x01, 0x15, '0', 0x0e, 0x02, 0x01, 0x01, 0x0a, 0x01, 0x00, 0xa1, 0x02, 0x82, 0x00, 0xa2, 0x02, 0x82, 0x00 ] As you can see, in addition to facility IE of len=30, we also have 30 zeros. Why libpri is sending this zeros? Examining add_callername_ie function, I discovered that it calls pri_call_apdu_queue only once. But in q931.c:send_message there are loop while (facevent) which iterates twice - adding normal PDU, then zeros string. Adding "break;" after "facevent->sent = 1;" fixes this issue, but it's a fix for symptom, not for the cause. | ||
Comments: | By: Russell Bryant (russell) 2005-10-03 14:04:04 Hey Matt! Can you take a look at this when you have time? By: Paul P Komkoff Jr (stingray) 2005-10-03 14:42:23 Here is some additional info based on today's experience. After I killed those mystical zeros, I still getting protocol error on setup from Samsung - but the call proceeding further. With zeros, I have got protocol error without any further processing. Maybe some other PBX interpreting those zeros more intelligent? By: Matthew Fredrickson (mattf) 2005-10-04 13:28:07 Your signaling type should be pri_cpe. Try changing it to that and see if still gives you this problem. By: Matthew Fredrickson (mattf) 2005-10-04 13:31:47 If there are still any problems, try to get a hold of me on IRC. My nick is Cresl1n. I should be on all day today. By: Matthew Fredrickson (mattf) 2005-10-04 13:37:47 I just found a little bug that might have caused this. Try updating libpri and see if it is fixed. By: Paul P Komkoff Jr (stingray) 2005-10-04 14:34:51 Yes it is fixed now. Thanks. I still have some questions though. I will ask you on IRC. Maybe you'll have some time to answer :) By: Matthew Fredrickson (mattf) 2005-10-04 14:43:54 Fixed in head |