From owner-freebsd-questions Mon May 8 18:14:21 1995 Return-Path: questions-owner Received: (from majordom@localhost) by freefall.cdrom.com (8.6.10/8.6.6) id SAA28200 for questions-outgoing; Mon, 8 May 1995 18:14:21 -0700 Received: from cs.weber.edu (cs.weber.edu [137.190.16.16]) by freefall.cdrom.com (8.6.10/8.6.6) with SMTP id SAA28176 for ; Mon, 8 May 1995 18:14:13 -0700 Received: by cs.weber.edu (4.1/SMI-4.1.1) id AA05256; Mon, 8 May 95 19:05:02 MDT From: terry@cs.weber.edu (Terry Lambert) Message-Id: <9505090105.AA05256@cs.weber.edu> Subject: Re: Streams Drivers? To: julian@ref.tfs.com (Julian Elischer) Date: Mon, 8 May 95 19:05:01 MDT Cc: marc@cmc.eng.comsat.com, questions@FreeBSD.org In-Reply-To: <199505081919.MAA14509@ref.tfs.com> from "Julian Elischer" at May 8, 95 12:19:21 pm X-Mailer: ELM [version 2.4dev PL52] Sender: questions-owner@FreeBSD.org Precedence: bulk > no it's not to SYSVish, but in the same breath, I must say I don't know > of any work being done on it either.. > karels has something called bstreams he's been mumbling about for a while, > and jolitz has something called 'flows' he's been ruminating on for about > as long.. terry MIGHT have a framework hidden away somewhere... I do; it's an SVR4.2 MP safe streams implementation. The bad news is that it runs only on 386BSD 0.1 + patchkit 1 and uses SVR4.0.2 sources for the ldterm and card drivers and uses Lachman sources for the TCP/IP. The other bad news is it panics when you use priority banding. The real bad news is that I finished it up after June of 1993, so Novell thinks they own it (they're wrong, but it takes a lot of strength to argue with the former USL). It would require a lot of work to get the pre-June '93 code up to snuff; I have it on a QIC-150 at home (without the USL/Lachman code it needs to link); the only other place it exists is on a 300M ESDI drive up at Weber (one that won't run because the WD1007 it was attached to is fried). Streams itself is almost trivial to write (well, excepting the priority banding stuff, obviously), with the big chunks going into the system call interface, the queue code, and the streams modules themselves (oh yeah; and the streams tail code to let them talk to drivers). Basically, the new networking code, the new syscall interface, and various other changes mean it's close to a total rewrite to make it go. On the other hand, the University of Arizona has Novell's SPX and IPX and their own TCP/IP running in an xkernel implementation, which is a very Streams-like environment. It might be worth grabbing that, or getting bstreams (if you can). Finally, there's a Streams implementation being done for Linux right now; I could dig out the reference if i had to... Terry Lambert terry@cs.weber.edu --- Any opinions in this posting are my own and not those of my present or previous employers.