From owner-freebsd-hackers Mon Jan 29 11:27:23 1996 Return-Path: owner-hackers Received: (from root@localhost) by freefall.freebsd.org (8.7.3/8.7.3) id LAA24057 for hackers-outgoing; Mon, 29 Jan 1996 11:27:23 -0800 (PST) Received: from phaeton.artisoft.com (phaeton.Artisoft.COM [198.17.250.211]) by freefall.freebsd.org (8.7.3/8.7.3) with SMTP id LAA24052 for ; Mon, 29 Jan 1996 11:27:19 -0800 (PST) Received: (from terry@localhost) by phaeton.artisoft.com (8.6.11/8.6.9) id MAA04239; Mon, 29 Jan 1996 12:24:10 -0700 From: Terry Lambert Message-Id: <199601291924.MAA04239@phaeton.artisoft.com> Subject: Re: NetBUI and/or IPX routing? To: dennis@etinc.com (dennis) Date: Mon, 29 Jan 1996 12:24:09 -0700 (MST) Cc: terry@lambert.org, hackers@freebsd.org In-Reply-To: <199601271732.MAA07882@etinc.com> from "dennis" at Jan 27, 96 12:32:49 pm X-Mailer: ELM [version 2.4 PL24] MIME-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7bit Sender: owner-hackers@freebsd.org Precedence: bulk > >The GetNearestServer request is a broadcast NCP. It is used by client > >machines starting up to locate a server on the local wire. For a proxy > >response, this would be the machine in the local bindery (SAP broadcast > >information is stored as tempory bindery objects in the local server's > >bindery -- under 4.x, this is done with bindery emulation as a local > >to each server NDS object). > > This is not generally true. A true IPX router maintains a server table > and a routing table and responds to requests directly. All such > requests are handled locally, and traffic is only routed once a > destination network number has been determined. I wouldn't call them > "proxies" per se....they are simply maintaining a map no unlike > some high level routing protocols for the IP world. Well, the one I worked on for Novell... 8-). A router won't respond to NCP requests, like GetNearestServer because it doesn't do NCP. It's not supposed to forward GetNearestServer requests because you don't want every machine up to 16 hops out seeing the request, so you set the hop count limit in the packet. GetNearestServer and SAP are not strictly related. SAP supplies the information to the bindery of servers and servers-acting-as-routers, and only servers-acting-as-servers handle NCP requests. In the case of a GetNearestServer for a remote-reset situation, the server that answers must have a disk image on the server to use as a pseudo "drive A:". Any server that can respond must have such an image, thus if a router claims it is the nearest server, then it must be a full server and not just a router. There are a number of the early "remote-reset" ROMs (the non-Ethernet II ROMs for Ne2000, for instance) that incorrectly looked at the source rather than the response address to get the response. A router doing a "GetNearestServer" response must diddle the source address in the packet for these clients to be happy. Now that said, I'm not saying that the packet forwarding in the IPX routing is not such that it will cause it to work (with potential noise overhead). Your best bet would be to contact Marty (I don't know where he works now) or Jim. Jim (Freeman) worked on the stack code that went into the NWU 4.x product that I worked on at Novell, and is a serious stack guru. His and Marty's code ended up being the baseline code for the UnixWare 2.x IPX code (Marty did the support for the link state routing). Jim Freeman currently works at Caldera (yes, the Linux place). Terry Lambert terry@lambert.org --- Any opinions in this posting are my own and not those of my present or previous employers.