[Asterisk-Users] Re: Bristuffed Asterisk: Hangup problems

[Marcel van der Boom]

Jeroen Zwarts wrote:
> 
> The only really strange thing I find in the logs that might have to do
> something with this is the following line on the verbosed console:
> chan_zap.c:8498 zt_pri_error: 1 updating callstate, peercallstate 2 to 1
> 
We had the exact same problem. It started happening for us starting at 
the 'k' release of bristuff (i mailed a msg on it in february i think 
to junghanns).

So, the 'i' release worked fine, while 'k' has the problem as described.

A quick diff of 'i' vs. 'k' showed me this (among other things):

diff -U0 -r -x '*.o' -x '*.so' 
bristuff-0.3.0-PRE-1i/libpri-1.2.2/q931.c 
bristuff-0.3.0-PRE-1k/libpri-1.2.2/q931.c
--- bristuff-0.3.0-PRE-1i/libpri-1.2.2/q931.c   2006-05-17 
19:54:51.000000000 +0200
+++ bristuff-0.3.0-PRE-1k/libpri-1.2.2/q931.c   2006-05-17 
20:04:07.000000000 +0200
@@ -4428,3 +4428,3 @@
-               if (c->ourcallstate != c->sugcallstate) {
-                   pri_error(pri, "updating callstate, ourcallstate 
%d to %d\n", c->ourcallstate, c->sugcallstate);
-                   c->ourcallstate = c->sugcallstate;
+               if (c->peercallstate != c->sugcallstate) {
+                   pri_error(pri, "updating callstate, peercallstate 
%d to %d\n", c->peercallstate, c->sugcallstate);
+                   c->peercallstate = c->sugcallstate;

This was such a close match, that i reversed that change in the 'k' 
release and voila! problem disappeared.
Now, i have no clue what kind of side-effects this has, if any, nor if 
this is the proper solution, but it made the problem disappear for us.

I haven't tried to apply the same to later bristuff releases (all 
releases up to 'p' give us the same hangup problem)

Hope this helps.

marcel