[Guy Harris <guy@xxxxxxxxxx>]

On Mon, Feb 03, 2003 at 12:12:30PM +0200, Teemu Rinta-aho wrote:
> Yes I am. I am printing the value in packet-ipv6.c and in
> packet-mip6.c and it's null already in IPv6 if the packet
> carries MIPv6. All other IPv6 protocols seem to have 
> non-null pinfo->cinfo. I have no idea where else I could
> look into.

You'll have to debug this one yourself; there's no code that a search
for "cinfo", followed by 0 or more tabs or spaces, followed by "=",
found that explicitly clears the "cinfo" field of the packet info
structure; there's only

	edt->pi.cinfo = cinfo;

in the beginning of "dissect_packet()", and that sets "edt->pi.cinfo" (a
pointer to "edt->pi" is passed as "pinfo" to "dissect_frame()", which
passes it on to subdissectors) to the value passed as an argument.

So either

	1) some piece of code is stomping on "pinfo->cinfo"

or

	2) it's null all the way back to the top-level dissector.