[Andy Ling <Andy.Ling@xxxxxxxxx>]

Thanks for the 2.0 info. That will be a background task.

 

Back to the immediate problem of decoding GIOP replies. Looking at 1.12 & 2.0 code the problem looks to be the same. It is the bit of code that works out a MFN (matching frame number?) from a FN and request ID. This is done by get_mfn_from_fn_and_reqid. A comment in this function says….

 

  /* Loop back from current end of complete_request_list looking for */

  /* a FN with the same reqid -- TODO enhance with port/address checks -- FS */

 

The TODO is the bit I’m missing.

 

So far as I can see nothing in this function has access to the IP addresses and port numbers of the frames being checked.

 

Presumably this would require comp_req_list_entry extending to include src & dest (or maybe just dest) addresses and port numbers and the get_mfn function extending to take the address and port number of the request being searched for.

 

I assume this information is the src/dst/srcport/dstport entries in the packet_info structure.

 

Any thoughts? Does this look like the way forward.

 

Thanks for any help

 

Regards

 

Andy Ling

 

 

From: wireshark-dev-bounces@xxxxxxxxxxxxx [mailto:wireshark-dev-bounces@xxxxxxxxxxxxx] On Behalf Of Graham Bloice
Sent: Mon 02 November 2015 12:37
To: Developer support list for Wireshark
Subject: Re: [Wireshark-dev] GIOP dissector reply decode

 

The Windows build for 2.0 has moved to CMake.  I'm in the process of submitting a change for the Developers Guide docs, but for now README.cmake in the top level source tree should have enough to get you going,

 

Nmake is still kind of supported, but nor for long.  The Win32.mak issue is detailed in the Developers Guide setup: https://www.wireshark.org/docs/wsdg_html_chunked/ChSetupWin32.html#ChSetupMSVC

 

On 2 November 2015 at 12:10, Andy Ling <Andy.Ling@xxxxxxxxx> wrote:

> Couple of thoughts from a quick skim of the code/git history:
 
> 1. I presume giop_complete_request_list is for matching request/reply.
>  Grepping its use will probably give you some clues as to where to look.

Thanks for the pointer. Rummaging in packet-giop.c it looks like requests and replies are stored
in a hash table giop_complete_reply_hash and the key used for this is packet_info->fd->num

Does that look right, am I on the right track?

If so, then I guess I need to look at where this number comes from and if it is
used for anything else. Any help anyone?

> 2. I'd recommend using the latest dev branch (master or at least master-2.0).
>  Any fixes you find would need to be applied there (master) first and then backported to 1.12 (and 2.0)

I had a quick look at 2.0. So far I've fallen over at the first hurdle. I tried following
the build instructions and the first nmake -f Makefile.nmake verify_tools fails
with a file win32.mak not found error. So I guess I'm looking at the wrong instructions.

Can someone point me at how to build 2.0 for Windows.

Thanks

Andy Ling

 

--

Graham Bloice

---

This email has been scanned for email related threats and delivered safely by Mimecast.
For more information please visit http://www.mimecast.com

---