From owner-freebsd-hackers  Mon Sep  8 18:44:04 1997
Return-Path: <owner-freebsd-hackers>
Received: (from root@localhost)
          by hub.freebsd.org (8.8.7/8.8.7) id SAA23191
          for hackers-outgoing; Mon, 8 Sep 1997 18:44:04 -0700 (PDT)
Received: from wong.rogerswave.ca (wong.rogerswave.ca [204.92.17.32])
          by hub.freebsd.org (8.8.7/8.8.7) with ESMTP id SAA23113
          for <freebsd-hackers@FreeBSD.ORG>; Mon, 8 Sep 1997 18:43:47 -0700 (PDT)
Received: (from wong@localhost) by wong.rogerswave.ca (8.8.7/8.7.3) id UAA00416; Mon, 8 Sep 1997 20:41:25 -0400 (EDT)
Date: Mon, 8 Sep 1997 20:41:22 -0400 (EDT)
From: Ken Wong <wong@a17b32.rogerswave.ca>
X-Sender: wong@wong.rogerswave.ca
Reply-To: wong@rogerswave.ca
To: "Kevin P. Neal" <kpneal@pobox.com>
cc: Terry Lambert <terry@lambert.olg>, freebsd-hackers@FreeBSD.ORG
Subject: Re: Divert sockets..
In-Reply-To: <1.5.4.32.19970907223328.008be880@mail.mindspring.com>
Message-ID: <Pine.BSF.3.91.970908202909.399A-100000@wong.rogerswave.ca>
MIME-Version: 1.0
Content-Type: TEXT/PLAIN; charset=US-ASCII
Sender: owner-freebsd-hackers@FreeBSD.ORG
X-Loop: FreeBSD.org
Precedence: bulk



On Sun, 7 Sep 1997, Kevin P. Neal wrote:
> Well, the Amiga OS had a system for message transferral. You would set
> up a message port. Do stuff. Then when you have nothing to do you sleep,
> waiting on a signal.
> 
> When you get a signal (because you got a message at one of your message
> ports) then you wake up and check the message.
> 
> The GUI notified you of events via just this messaging scheme. ARexx
> programs could direct you to do stuff through another message port. This was
> so common that most programs main loop's were just a simple 
> 
> Wait()
> 
> Get Message()
> 
> Copy Data Out of message
> 
> Reply Message()
> 
> Do Stuff()
> 
> Loop back to Wait().
> 
> The File Notification scheme (useful for Terry's file browser) may have used
> this messaging scheme as well (I honestly don't remember). You certainly
> could be notified of disk change events (floppy insert/removal, later
> removable HD change) through messages.
> 
> Now, here's the question: How difficult would it be to do something like this?
> It would take some sort of shared memory or other scheme to get the data from
> one program to another. Or would that be too general? How about just a scheme
> to get messages to and from the kernel? 

I have a kernel patch do just that. actually, it applys to alot of OS;
QNX, Tandem Non-Stop Kernel, Minix.

I don't know how if possible to include/commit it in the kernel tree.
anyhow, If you are interested, I can send you the files.

it is basically can do the following:

prog A:

main()
{
	while(1) {
	pid = receive(pid-mask, buf); // blocked
        ...
        reply(pid, buf);
	}
}

prog B:
main()
{
	send(pid, buf); // blocked
}