From owner-freebsd-hackers Thu Jan 11 10:13:57 1996 Return-Path: owner-hackers Received: (from root@localhost) by freefall.freebsd.org (8.7.3/8.7.3) id KAA08229 for hackers-outgoing; Thu, 11 Jan 1996 10:13:57 -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 KAA08222 for ; Thu, 11 Jan 1996 10:13:52 -0800 (PST) Received: (from terry@localhost) by phaeton.artisoft.com (8.6.11/8.6.9) id KAA17651; Thu, 11 Jan 1996 10:53:20 -0700 From: Terry Lambert Message-Id: <199601111753.KAA17651@phaeton.artisoft.com> Subject: Re: samba and win95 To: joerg_wunsch@uriah.heep.sax.de Date: Thu, 11 Jan 1996 10:53:20 -0700 (MST) Cc: freebsd-hackers@FreeBSD.org In-Reply-To: <199601110905.KAA11027@uriah.heep.sax.de> from "J Wunsch" at Jan 11, 96 10:05:56 am 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 > > As far as I know, it isn't. smbclient is fine to access files "by hand" > > or crash WNT servers. That's all. You can't mount a SMB volume > > under Unix. > > No, not really `mount' in the Unix sense of `mount a file system' > (unless you're running Linux :). But you can get access to those > files, even a recursive `get' operation is available. If you are using Linux's SMBFS, it's not "in the Unix sense", it's in the Linux sense. In particular, there are issues about when credentials are evaluated. For NetWare, LanMan, etc., credentials are evaluated on connect, where the machine authenticates to the server as a single user. There is no way, short of implementing a VMS style broadcast mechanism and a finite state automaton for the terminals (every wonder why Wyse 50's suck for VMS?) so that broadcast delivery is blocked unless the automaton is at ground state. Even then, you are potentially screwed by certain issues, unless you force the automaton (and thus the terminal) back to ground state after a broadcast including escape sequences. You need the broadcast mechanism to allow the kernel to ask each user for authentication information. Your "client" will need one "netname" per user contacting each server, and you will probably need aliasing to distinguish inbound packets, which are sent to hardware address after the cache lookup for the request. This *could* work on NetWare if you have a NetWare 3.x orr above only and used an internal network with a pseudo node per client (like in the NUC product). Terry Lambert terry@lambert.org --- Any opinions in this posting are my own and not those of my present or previous employers.