RE: [Ethereal-dev] MEGACO plugin seriously flawed?

["Anders Broman (AL/EAB)" <anders.broman@xxxxxxxxxxxx>]

Hi,
I checked in similar changes a few days ago:
http://anonsvn.ethereal.com/viewcvs/viewcvs.py/trunk/plugins/megaco/packet-megaco.c?rev=12043&r1=11997&r2=12043
Could you check it out and see if it works ok ?

Best regards
Anders

-----Original Message-----
From: ethereal-dev-bounces@xxxxxxxxxxxx
[mailto:ethereal-dev-bounces@xxxxxxxxxxxx]On Behalf Of Ruud Linders
Sent: den 25 september 2004 12:54
To: Ariel Burbaickij; Ethereal development
Subject: Re: [Ethereal-dev] MEGACO plugin seriously flawed?



Ariel et al,

I didn't know about SCTP NOT requiring a TPKT header but TCP indeed
DOES require a TPKT encapsulation.

Several months ago I made a quick-and-dirty patch to allow for TPKT
encapsulation and de-segmentation which is working fine for me.
However, it needs some cleaning up
- adding preference to be able to turn off/on de-segmentation,
   the global var is already there.
- correct check in the dissect_megaco_tpkt routine for the MEGACO
   string (for whatever reason I only checked for 'M' 'E' (probably
   lazines at the time!)
- proper skip/check for whitespace in same check
- registration/hook for sctp support

Attached the patch against 0.10.6, feel free to use it.

Regards,
	Ruud

Ariel Burbaickij wrote:
> Hello dear mailing list participants,
> 
> following code exceprt from MEGACO plugin:
>   /*
>      * Check to see whether we're really dealing with MEGACO by looking
>      * for the MEGACO string.  This needs to be improved when supporting
>      * binary encodings.
>      */
>     if(!tvb_get_nstringz0(tvb,0,6,word)) return;
>     if (strncasecmp(word, "MEGACO", 6) != 0) return;
> 
> This is incorrect as was found out today. IIUC what is going one this code
> omits the possbility of TPKT header between TCP and MEGACO which
> is reuqired as per RFC 3015 Addendum D. TPKT (RFC 1006) caters for message
> segmentation. So either documentation needs to be updated stating
> that MEGACO works only with SCTP where there is no such thing
> as message segmentation or code needs to be corrected. Would
> be glad to hear your opinion.
> 
> With Best Regards
> Ariel Burbaickij
> 
> _______________________________________________
> Ethereal-dev mailing list
> Ethereal-dev@xxxxxxxxxxxx
> http://www.ethereal.com/mailman/listinfo/ethereal-dev