A little progress on this problem

Examining the logs i found a weird looking 'soft hangup' which reminded me on an earlier issue we had. (and the reason why we were still on the 'i' release of bristuff). It looked as if the channel hung up just before rxfax actually could begin to work.

Normally we would let the faxdetection be automatic and let ast jump to the 'fax' extension, which in turn calls the faxreceive macro as described in my original post.

Bypassing all that and directly tie-ing an extension to the faxreceive macro (not even answering the channel), gives a successful fax reception. RxFax has a check whether the channel is answered, and it will answer it, if not already.

I think this was a lucky shot in the dark though. The problem seems to be a mismatch between the state of the channel (answered) and what it actually is. Getting data from the channel wont work then i guess.

This is a suitable workaround for our little setup for now. The only thing we miss at this point is that we wont be able to receive faxes at every extension anymore, just the one.

I do not have enough knowledge of the sources to suggest a fix for this. It looks like either the specific stuff for our card (quadbri) or asterisk itself would be the area to look into, but again, i am not (yet) capable of doing so myself.

Hope this helps anyone fixing the real problem.

marcel

On 16 nov 2006, at 11:27, Marcel van der Boom wrote:

Hi,

I'm using spandsp-0.0.3
[http://www.soft-switch.org/downloads/snapshots/spandsp/ spandsp-20061116.tar.gz]

on a bristuffed asterisk (1.2.13)
[http://212.91.251.199/~junghanns.net/downloads/bristuff-0.3.0- PRE-1v.tar.gz]

libtiff is at version 3.6.0

Running on: Linux router2 2.6.17-2-686 #1 SMP Wed Sep 13 16:34:10 UTC 2006 i686 GNU/Linux
Debian testing distro.

I've tried many combinations of bristuffed ast and spandsp versions, but all fail at the same point. The last combination i got to work was bristuffed 0.3.0-PRE-1i with spandsp-0.0.2-pre25 (on an earlier kernel)

The app_rxfax.c in use is from:
[http://www.soft-switch.org/downloads/snapshots/spandsp/test-apps- asterisk-1.2/app_rxfax.c]

On reception of a fax through RxFax, i get the exception. The relevant part of the dialplan is

[macro-faxreceive]
; Receive a fax
exten => s,1,Set(FAXFILE=${FAXSPOOL}/${UNIQUEID}.tif) ; Save the fax in a tif file
exten => s,2,RxFAX(${FAXFILE})                          ; Receive it
exten => s,3,NoOp("Fax reception complete")             ;
exten => s,4,Hangup

Running asterisk (with the above versions) through gdb and doing a backtrace gives me:

#0  0xa7d45947 in raise () from /lib/tls/libc.so.6
#1  0xa7d470c9 in abort () from /lib/tls/libc.so.6
#2  0xa7d7afda in __fsetlocking () from /lib/tls/libc.so.6
#3  0xa7d8289f in mallopt () from /lib/tls/libc.so      .6
#4  0xa7d82942 in free () from /lib/tls/libc.so.6
#5 0xa75efd68 in rxfax_exec (chan=0x818c5f8, data=0xa74a4798) at app_rxfax.c:327 #6 0x08090088 in pbx_extension_helper (c=0x818c5f8, con=<value optimized out>, context=<value optimized out>, exten=0x818c83c "s", priority=2,
    label=0x0, callerid=0x0, action=1) at pbx.c:554
#7 0xa762cb05 in macro_exec (chan=0x818c5f8, data=0xa74aafe8) at app_macro.c:221 #8 0x08090088 in pbx_extension_helper (c=0x818c5f8, con=<value optimized out>, context=<value optimized out>, exten=0x818c83c "s", priority=1,
    label=0x0, callerid=0x0, action=1) at pbx.c:554
#9  0x08091dee in __ast_pbx_run (c=0x818c5f8) at pbx.c:2231
#10 0x08092a1c in pbx_thread (data=0x818c5f8) at pbx.c:2518
#11 0xa7f0d0bd in start_thread () from /lib/tls/libpthread.so.0
#12 0xa7de892e in clone () from /lib/tls/libc.so.6

This seems to indicate that the ast_frfree(inf); at line 327 of app_rxfax.c causes the problem chain?

I'm a bit lost on how to debug this further. Is this actually a spandsp problem or is another package the cause?
Any tips?

marcel


--
Marcel van der Boom
HS-Development BV               --   http://www.hsdev.com
So! webapplicatie framework  --   http://make-it-so.info


_______________________________________________
--Bandwidth and Colocation provided by Easynews.com --

asterisk-users mailing list
To UNSUBSCRIBE or update options visit:
   http://lists.digium.com/mailman/listinfo/asterisk-users

--
Marcel van der Boom
HS-Development BV               --   http://www.hsdev.com
So! webapplicatie framework  --   http://make-it-so.info


_______________________________________________
--Bandwidth and Colocation provided by Easynews.com --

asterisk-users mailing list
To UNSUBSCRIBE or update options visit:
  http://lists.digium.com/mailman/listinfo/asterisk-users

Reply via email to