Ethereal-dev: Re: [Ethereal-dev] Help with extending RSVP dissector

Note: This archive is from the project's previous web site, ethereal.com. This list is no longer active.

From: srinivas reddy <srinivasreddy_m@xxxxxxxxx>
Date: Mon, 23 Jan 2006 10:32:35 -0800 (PST)

--- Ashok Narayanan <ashokn@xxxxxxxxx> wrote:

> Is this different from the way in which objects are
> handled in RSVP?
> In the RSVP dissector I used loops to walk through
> the set of objects,
> and also the set of TLVs inside some objects (like
> ERO).
> 
> In the past I used opaque adds like
> proto_tree_add_guint() for many
> fields in RSVP. The two issues I have seen with this
> in the past are:
> 
> - Tree management: if you use the same tree index
> for all the
> elements, then opening a single subtree and then
> switching packets
> will results in all the subtrees being opened
> immediately. 

How do I solve this even if I register all the fields?
Ideally I want each filter to have its own subtree.
Since I don't know how many filters are going to be
sent, how can I define a number of tree indexes. I
guess one way is to use some large number.

> 
> - Filtering: you can't filter on an unnamed entity.

I guess this can be solved by registering all the
fields. But is this the only advantage of registering
fields?

> 
> Having said that, RSVP has a myriad of different
> object types and
> values, so if you don't want to name all of them and
> are willing to
> live with the issues above, feel free.
> 
> -Ashok
> 
> On Fri, Jan 20, 2006 at 11:51:58PM -0800, srinivas
> reddy wrote:
> > Hi,
> > I am new to Ethereal development and need some
> help in
> > extending RSVP dissector. RSVP is extended in
> > non-standard compliant way to support QoS in CDMA
> > network and a new object is added to send filter
> > specification to the network. A filter
> specification
> > may contain one or more filters. Each filter may
> have
> > fields like source address, destination address,
> TOS,
> > port numbers etc.
> > 
> > I saw in readme.developer file that each field in
> a
> > protocol needs to be registered using an array of
> > hf_register_info structures. In this case, since
> there
> > could be more than one set of repetetive fields,
> how
> > do I register them? Is it ok if I don't register
> them
> > and parse the fields myself and add them to
> protocol
> > tree using proto_tree_add_guint*().
> > 
> > Your help is greatly appreciated.
> > 
> > Thx,
> > Srinivas
> > 
> > 
> > 
> > Its what inside that makes the difference -- Lance
> Armstrong
> > 
> > __________________________________________________
> > Do You Yahoo!?
> > Tired of spam?  Yahoo! Mail has the best spam
> protection around 
> > http://mail.yahoo.com 
> > 
> > _______________________________________________
> > Ethereal-dev mailing list
> > Ethereal-dev@xxxxxxxxxxxx
> >
>
http://www.ethereal.com/mailman/listinfo/ethereal-dev
> 
> -- 
> 
> 
> 
> --- Asok the Intern
> ----------------------------------------
> Ashok Narayanan
> IOS Network Protocols, Cisco Systems
> 1414 Mass Ave, Boxborough MA 01719
> Ph: 978-936-1608.  Fax: 978-936-2218 (Attn: Ashok
> Narayanan)
> 
> _______________________________________________
> Ethereal-dev mailing list
> Ethereal-dev@xxxxxxxxxxxx
>
http://www.ethereal.com/mailman/listinfo/ethereal-dev
> 


Its what inside that makes the difference -- Lance Armstrong

__________________________________________________
Do You Yahoo!?
Tired of spam?  Yahoo! Mail has the best spam protection around 
http://mail.yahoo.com