[asterisk-dev] Segfaults with 1.4.0 and E1s

[Gavin Hamill]

On Saturday 30 December 2006 22:19, Matthew Fredrickson wrote:
> Can you post a detailed bug report on bugs.digium.com as well as a
> complete backtrace?  

Sure thing. 

> My first thought is that the id value is very 
> wrong.  In fact, it shouldn't even get to this line of code with the
> current version of libpri since the id is a negative value.  Are you
> sure you updated libpri and zaptel as well as asterisk when you updated
> to 1.4?

Yes, libpri + zaptel were both updated before the new 1.4.0 was started. I've 
never used a pre-packaged version or manually copied libpri+zaptel.. only 
ever tar xzvf.. (configure) .. make .. make install...

'locate libpri' only returns the source code in /usr/src, and these 4 files 
in /usr/lib:

hash:/usr/src/asterisk-1.4.0# ls -l /usr/lib/libpri.*
-rw-r--r-- 1 root root 427240 2006-12-28 20:13 /usr/lib/libpri.a
lrwxrwxrwx 1 root root     13 2006-12-28 20:13 /usr/lib/libpri.so -> 
libpri.so.1.0
lrwxrwxrwx 1 root root     13 2006-12-28 20:13 /usr/lib/libpri.so.1 -> 
libpri.so.1.0
-rwxr-xr-x 1 root root 307191 2006-12-28 20:13 /usr/lib/libpri.so.1.0

And those date/timestamps match up perfectly with the same filenames 
in /usr/src/libpri-1.4.0.

-rw-r--r-- 1 root root 625118 2006-12-28 
20:23 /lib/modules/2.6.18-3-amd64/misc/zaptel.ko

this matches up with the compiled source in /usr/src/zaptel-1.4.0

The only slightly different thing I'm doing here is using a Sangoma A104 which 
has a tiny patch to zaptel to support hardware D-Channel processing:

http://lists.alioth.debian.org/pipermail/pkg-voip-commits/2006-November/001828.html

Cheers,
Gavin.