Wireshark-dev: Re: [Wireshark-dev] Infiniband dissector not showing payload
From: Alexis La Goutte <alexis.lagoutte@xxxxxxxxx>
Date: Fri, 21 Mar 2014 17:05:25 +0100
Hi David,

A good idea, it is to create a bug in bugtracker (
https://bugs.wireshark.org ) and attach your pcap sample.

Regards,

On Thu, Mar 20, 2014 at 4:07 PM, David Ameiss <netshark@xxxxxxxxxxxxx> wrote:
> In trying to look at some IB captures, I noticed that any payload data
> included was never displayed in the packet details (tree) pane. It obviously
> appears in the bytes pane - but nothing in the details pane. This is with
> both 1.10.6 and the latest git master.
>
> I did some poking around and debugging, and it appears that the "data"
> dissector is always called with a NULL tree.
>
> Now, I'm not that familiar with IB. However, I was able to get the payload
> data displayed by making a change in epan/dissectors/packet-infiniband.c, in
> dissect_infiniband_common(), changing
>
>     /* Get the parent tree from the ERF dissector.  We don't want to nest
> under ERF */
>     if (tree && tree->parent)
>     {
>         /* Set the normal tree outside of ERF */
>         tree = tree->parent;
>         /* Set a global reference for nested protocols */
>         top_tree = tree;
>     }
>
> to
>
>     /* Get the parent tree from the ERF dissector.  We don't want to nest
> under ERF */
>     if (tree && tree->parent)
>     {
>         /* Set the normal tree outside of ERF */
>         tree = tree->parent;
>         /* Set a global reference for nested protocols */
>         top_tree = tree;
>     }
>     else
>     {
>         top_tree = tree;
>     }
>
> Debugging showed that tree->parent was always NULL. Making this change
> caused the Data to be displayed at the same tree level as Frame, Extensible
> Record Format, and Infiniband. Not sure this is correct, since the data is
> actually within the Infiniband subtree.
>
> I can provide more details, but I don't know what other impact the above
> change might have. I was hoping someone more familiar with the IB dissector
> might determine what the proper fix is.
>
> --
> David Ameiss
> netshark@xxxxxxxxxxxxx
> ___________________________________________________________________________
> Sent via:    Wireshark-dev mailing list <wireshark-dev@xxxxxxxxxxxxx>
> Archives:    http://www.wireshark.org/lists/wireshark-dev
> Unsubscribe: https://wireshark.org/mailman/options/wireshark-dev
>             mailto:wireshark-dev-request@xxxxxxxxxxxxx?subject=unsubscribe