RE: T68i as SMSC

Hello to all,

thank you Oded for the quick fix. 

To share the problem with the community:

I now found the problem, it was a (simple) misconfiguration 
of the SMSC: I added "no-smsc = true" (I've misinterpreted
the user manual), but actually there was the smsc phone
number in the PDU, so the at2_pdu_extract skipped
the removal of the smsc phone number and the decode_deliver_sm()
failed when looking at the TP-OA address.

Thanks,

Peter

 


-----Original Message-----
From: Oded Arbel [mailto:[email protected]]
Sent: Mittwoch, 6. November 2002 12:34
To: Schaich, Peter; Kannel-devel (E-mail)
Subject: RE: T68i as SMSC



I've added some safeguards into decode_deliver_sm() so that it will break
gracefully if the TP-OA address length field is too long, or it reaches the
end of the PDU too early.
Patch commited to CVS - its a bug fix, it resolves a crash, it works here
and it shouldn't break anything whatever you do.

--
Oded Arbel
m-Wise mobile solutions
[email protected]

+972-9-9581711 (116)
+972-67-340014

::..
I don't mind living in a man's world as long as I can be a woman in it.
                         - Marilyn Monroe


> -----Original Message-----
> From: Oded Arbel 
> Sent: Wednesday, November 06, 2002 12:51 PM
> To: Schaich, Peter; Kannel-devel (E-mail)
> Subject: RE: T68i as SMSC
> 
> 
> 
> 
> > -----Original Message-----
> > From: Schaich, Peter [mailto:[email protected]]
> > 
> > I now have tracked (debugged) the problem down to
> > the PDU type the T68i returns. It is not 0, but 3,
> 
> 3 is defined as "reserved". from 3GPP TS 23.040 :
> <snip>?
> In an MS receives a TPDU with "Reserved" value in the TP-MTI, 
> it shall process the message as if it were an SMS-DELIVER bu 
> store the message exactly as received.
> </snip>
> 
> I think that your problem is not with the T68i, but with your 
> network operator's SMSC which sends SMs with type "Reserved". 
> 
> > so the PDU is not decoded by the at2 smsc. I tried
> > to allow this pdu type, but then the decode fails
> > completely.
> 
> How did you try to decode it ?
> 
> > 2002-11-06 10:57:15 [6] DEBUG: AT2[m3s]: <--
> > 0791947122720000040C9194717206979900002011600145138004F4F29C0E
> OK , this is interesting - if we assume this message is a 
> DELIVER_SM, then the originating address field is set to be 
> an international MSISDN (which makes sense) but the length of 
> the address is set to be 94h bytes long - which is quite 
> literaly longer then the entire PDU. this is what breaks the 
> bearerbox. I'll try to fix this in the AT2 driver, but in any 
> case - this SM does not make much sense as a DELIVER_SM, so 
> it cannot be a legal SM according to the SPECS, and I don't 
> expect that Kannel should be able to parse it correctly.
> 
> --
> Oded Arbel
> m-Wise mobile solutions
> [email protected]
> 
> +972-9-9581711 (116)
> +972-67-340014
> 
> ::..
> Time is what you make of it.
>   -- Swatch slogan
> 
>