Date: Wed, 2 May 2001 15:55:09 -0600
Message-Id: <200105022155.f42Lt9Z30531@vindaloo.ras.ucalgary.ca>
From: Richard Gooch <rgooch@ras.ucalgary.ca>
To: BieshaarP@netscape.net (Peter Bieshaar)
Cc: netdev@oss.sgi.com
Subject: Re: possible suggestion for within the Linux kernel
In-Reply-To: <652B5AAB.43CD19F0.0225591B@netscape.net>
References: <17CEC8ED.094A580A.0225591B@netscape.net>
	<200105011632.f41GW9U12212@vindaloo.ras.ucalgary.ca>
	<652B5AAB.43CD19F0.0225591B@netscape.net>
Sender: owner-netdev@oss.sgi.com
Precedence: bulk

[Please fix your MUA to wrap your lines at 72 characters]
[Please quote first and put your response after the quoted material]

Peter Bieshaar writes:
> Richard Gooch <rgooch@ras.ucalgary.ca> wrote:
> >
> > Peter Bieshaar writes:
> > > I got this email address from Alan Cox. I think I have a really
> > > great idea to implement into the Linux kernel.
> > > But can actually not find any place to talk with anybody about
> > > this. It is about tcp_doors, a door equivalent like Solaris > 2.51
> > > is using but then built onto the tcp layer.
> > > 
> > > It will have great impact into a lot of OS stuff like mem mngmnt,
> > > IPC, function, security, a lot of other exceptions and stack
> > > management to name a few. I want to know what your ideas are and how
> > > something like this should be set up.
> > 
> > What's the benefit to having tcp_doors?
>
> I think the benefit will be very much in environments where a couple
> of little, probably embedded, CPU's can communicate with more
> intelligence and faster. I think about the car-industry and possibly
> droids. Both can be done with current technology but that will
> always be a three steps communication, with doors it should go more
> directly.

You didn't answer the question. You just talked about where it is
claimed to help. *How* will it help?

				Regards,

					Richard....
Permanent: rgooch@atnf.csiro.au
Current:   rgooch@ras.ucalgary.ca