Re: Some questions on RPC dissectors (for a new Gluster dissector)

[Kaul <mykaul () gmail com>]

On Fri, Apr 20, 2012 at 6:08 PM, Niels de Vos <ndevos () redhat com> wrote:

> Hi all,
>
> Bug 5773 was opened as an RFE for implementing a dissector for Gluster.
> The Gluster 'protocol' consists out of several RPC-programs, each with
> their own set of procedures.
>
> There are some questions I would like to ask:
>
> 1)
> I am not sure what the best practice is to group these related
> RPC-programs. At the moment, I have a packet-gluster.c file (and a header)
> that registers proto_gluster with proto_register_protocol().
>
> Each dissector for an RPC-program registers its header-field-array with
> this proto_gluster handle. Some of the RPC-programs have their own file, to
> the proto_gluster handle is made non-static and exported in the
> packet-gluster.h file.
>
> Is this something that I should rather not do, and create complete
> separate dissectors for each RPC-program? There are some header-fields that
> are the same for several RPC-programs, should these just be duplicated?
> (Currently these are non-static as well, and listed in the header file too.)


You can have a packet-gluster-common.{c|h} file(s).


> If I can keep using this structure, how can I guarantee that the
> proto_gluster handle has been initialized when other dissectors try to use
> it with proto_register_field_array()?
>
> 2)
> The Gluster protocols use RPC-credentials with number 5. This number is
> currently not dissected in packet-rpc.c, but I also doubt IANA assigned
> this number to the Gluster protocols. What would be the best way to add an
> implementation to dissect the credentials in the RPC-header?
>
> a) just dissect any number 5 flavour as Gluster-credentials

Google around to see if others may have (ab)used this number as well. If
not, I'd just dissect it as Gluster's - with a comment in the code
explaining it is non registered to Gluster (and worth shooting an email to
the gluster dev list asking if they intend to register it).

Y.

b) add a preference-option to the RPC-dissector
> c) detect the credential-flavour based on flavour+program number
> d) something else, please specify
>
>
> That's it for now. I'm happy to provide more details if that would clarify
> my questions.
>
> Many thanks,
> Niels
> ______________________________**______________________________**
> _______________
> Sent via:    Wireshark-dev mailing list <wireshark-dev () wireshark org>
> Archives:    http://www.wireshark.org/**lists/wireshark-dev<http://www.wireshark.org/lists/wireshark-dev>
> Unsubscribe: 
> https://wireshark.org/mailman/**options/wireshark-dev<https://wireshark.org/mailman/options/wireshark-dev>
>            mailto:wireshark-dev-request@**wireshark.org<wireshark-dev-request () wireshark org>
> ?subject=**unsubscribe
___________________________________________________________________________
Sent via:    Wireshark-dev mailing list <wireshark-dev () wireshark org>
Archives:    http://www.wireshark.org/lists/wireshark-dev
Unsubscribe: https://wireshark.org/mailman/options/wireshark-dev
             mailto:wireshark-dev-request () wireshark org?subject=unsubscribe