Re: Calling a dissector: Type for data parameter

[João Valverde via Wireshark-dev <wireshark-dev () wireshark org>]

On 16/06/21 15:36, David Perry wrote:
> Sorry to drag up an old topic, but I've been thinking about this:
>
> > Message: 5
> > Date: Sat, 29 May 2021 09:32:29 +0200
> > From: Anders Broman <a.broman58 () gmail com>
> > To: Developer support list for Wireshark <wireshark-dev () wireshark org>
> > Subject: Re: [Wireshark-dev] Calling a dissector: Type for data
> >     parameter
> > Message-ID:
> >     <CAOpyz=zDycm33PXUwtBCTew7gTTEcSLiJ-f8SHW0L-863Q517A () mail gmail com>
> > Content-Type: text/plain; charset="utf-8"
> >
> > Hi,
> > Yes the method is fragile. At the time of development I think it was
> > proposed to pass a struct containing a string and the void pointer where
> > the string could be used as a identifier. But that was voted down.
> > Regards
> > Anders
>
> I wasn't around for that discussion so I don't know the reasons, but 
> how does this sound as a refined approach?:
>
> * Define a `dissector_data_t` that has a `guint32` identifier field, 
> and a `void *` data field.
>
> * Replace the `void *data` parameter to dissectors with a pointer to a 
> `dissector_data_t`.
>
> * Either:
>
>     * Easy way: maintain a static list of identifiers that map to 
> expected data types, or
>
>     * Have dissector X request an identifier in its registration 
> function for the type of data it expects, and have dissector Y (which 
> will call X) request, in its handoff function, the identifier of the 
> type of data it needs to pass to X.
>
> * Dissectors check for the right identifier in their 
> `dissector_data_t` parameter and don't try to use it if it's wrong.
>
> Thoughts?

I think what you suggest would be the most straightforward fix.

To avoid breaking backward compatibility and changing thousands of 
dissectors at the same time, both of which are highly problematic, it 
can be done by adding a new dissector type (like it was done with 
"dissector_cb_t", only using a different signature).[1]

Also a giant static list of dissector_data_t type identifiers would be 
pretty clunky. I think we could recycle the protocol registration number 
for that.

João


[1]https://gitlab.com/wireshark/wireshark/-/blob/master/epan/packet.h#L79


> David
>
>
> ___________________________________________________________________________ 
>
> Sent via:    Wireshark-dev mailing list <wireshark-dev () wireshark org>
> Archives:    https://www.wireshark.org/lists/wireshark-dev
> Unsubscribe: https://www.wireshark.org/mailman/options/wireshark-dev
> mailto:wireshark-dev-request () wireshark org?subject=unsubscribe

___________________________________________________________________________
Sent via:    Wireshark-dev mailing list <wireshark-dev () wireshark org>
Archives:    https://www.wireshark.org/lists/wireshark-dev
Unsubscribe: https://www.wireshark.org/mailman/options/wireshark-dev
            mailto:wireshark-dev-request () wireshark org?subject=unsubscribe