[CUWiN] Question about Originating ifIndex field

[Sanjai Narain]

Hi David: Thanks a lot for this helpful information. Regards. -- Sanjai

David Young wrote:

> On Sun, Apr 17, 2005 at 07:58:48PM -0400, Sanjai Narain wrote:
> >
> > Is this field set correctly in the HSLS Packet Header? It refers to the
> > interface number in the SNMP mib, but if the host is not running SNMP then
> > how does one interpret this field? Does it correlate with any information
> > produced by Linux's ifconfig?
>
> Sanjai,
>
> The interface number is set in hsls_interface_create.  The call
> chain there is hsls_interface_create->getifprops->if_nametoindex.
> The daemon sets the originator interface in a couple places.
> One is in hsls_router_retransmit_lsupd.  The other place
> is hsls_interface_send_olsu, where the call chain is
> hsls_interface_send_olsu->hsls_interface_postpack_lsu.
>
> It is more than likely that the SNMP MIB and the operating system's
> interface index will be the same.  I may have made a mistake to refer
> to the SNMP MIB, but this seems to be the way that a lot of RFCs refer
> to an interface index in the abstract.  Suffice it to say, I expect
> for the ifindex of every physical interface to be different from every
> other physical interface's; I also expect for the ifindex to be greater
> than zero.
>
> I don't think the information correlates with ifconfig.  I believe
> that on *BSD, 'ifconfig -a' lists the interfaces in ifindex order,
> but I chalk that up to a happy accident of implementation.
>
> If all the ifindices are the same, it may be that each Linux box on
> your network has coincidentally assigned the same ifindex to its HSLS
> interface.  However, if any two interfaces on the same Linux box have
> the same ifindex, that's a bug.
>
> If it helps, NetBSD's if_nametoindex(3) manual page says that it the
> library routine conforms to an RFC:
>
>   STANDARDS
>        The if_nametoindex(), if_indextoname(), if_nameindex(), and
>        if_freenameindex() functions conform to RFC 2553.
>
> Dave
>
> >
> > In the packet traces that I have looked at, this number is always set to
> > 4.
> >
> > Thanks for any help. Regards. -- Sanjai
> > _______________________________________________
> > CU-Wireless mailing list
> > CU-Wireless at lists.cuwireless.net
> > http://lists.chambana.net/cgi-bin/listinfo/cu-wireless
> > Project Page: http://cuwireless.ucimc.org
>
> --
> David Young             OJC Technologies
> dyoung at ojctech.com      Urbana, IL * (217) 278-3933