[Asterisk-Users] one-way audio and isdn4linux

[andre]

Hi,

Apologies in advance for the lengthy email.

I'm new to asterisk and have trouble with isdn4linux.

The setup is very basic like this:

   winxp ------- asterisk -------- winxp
   x-lite           |              x-lite
                    |
                   pstn

The hardware involved is:
Compaq EVO with RH9/kernel 2.4.20-30.9.
Fritz!Card PCI v2
Asterisk CVS-04/17/04-21:36:18

Basically I run:
modprobe -a hisax_fcpcipnp
before I start asterisk with -vvvvvgc

The configs are like this:
MODEM.CONF:
[interfaces]
context=default
driver=aopen
driver=i4l
language=en
type=autodetect
stripmsd=0
dialtype=tone
mode=immediate
device => /dev/ttyI0
device => /dev/ttyI1
group=1
msn=0208080808
incomingmsn=*
device => /dev/ttyI0
device => /dev/ttyI1

EXTENSIONS.CONF:
[general]
static=yes
writeprotect=no
[globals]
TRUNK=Modem/g1
TRUNKMSD=0
[default]
exten => _9001,1,Dial(${TRUNK}:0208080809,60,tr)

Firewalls are disabled on all machines (doublechecked).
Sip calls between the two x-lite clients work consistently.
After I boot the asterisk server I can make one call succesfull to the
pstn (two-way voice). Subsequent calls from x-lite to pstn only give me
one-way voice; the direction that works is pstn->sip: a call is
cleary established, I just cannot understand why voice is lacking
sip->pstn or why it works the first time after a reboot.

I had a look on the sip-clients and on the server with ethereal, and
two-way traffic is happening consistently.
Show channels within the CLI saw 'frames in' and 'frames out' augmenting
on all channels.
Looking at /dev/isdnctrl I see slightly different isdn messages, but in
both cases I got a: L3DC ChangeState ST_L3_LC_ESTAB

I have tried to compile chan_capi but ran into the same issue as
described on a posting on april 12th:
[Asterisk-Users] Trouble compiling chan_capi on Suse 9.0

As I had two servers and two fritz cards I tried both with identical
software and both behave the same.

I hope I have made a stupid mistake and somebody on the list can point
it out to me. Every suggestion is of course most welcome.

I'm a bit unsure about the machines though: both machines crashed one or
more times completely (froze up) during a call to the pstn and dumped
core. Basically a hardware reset was all that I could do. One of these
times I was looking at the console and wrote down this:

Code: 8b 38 8b 58 04 8b 68 10 c7 44 24 18 00 00 00 00 ff 4c 24 40
 <0>Kernel Panic: Aiee, killing interrupt handler!
 In interrupt handler - not syncing

Thanks, andre