> From: Jason Thorpe <thorpej@nas.nasa.gov>
> Date: 1999-09-25 11:39:14 -0700
> To: Ignatios Souvatzis <ignatios@cs.uni-bonn.de>
> Subject: Re: How does 10baseT/FDX work?
> Cc: tech-net@netbsd.org
> Delivered-to: tech-net@netbsd.org
>
> On Sat, 25 Sep 1999 19:19:07 +0200
>  Ignatios Souvatzis <ignatios@cs.uni-bonn.de> wrote:
>
>  > the i82595TX does not have any official FDX support. However,  
it has an
>  > "external loopback" mode, where it can transmit and receive  
simultaneously.
>  >
>  > Is this enough to do 10baseT-FDX, as long as I don't want  
autonegotiation,
>  > which I guess can't work with this?
>
> Yes, that should be sufficient to do 10baseT-FDX.  There's nothing  
really
> magical about full-duplex mode.
>
> Hopefully, the chip turns off collision detection when you enable this 
> feature (I would assume it does).

It may work; I don't know exactly what external loopback does in  
most devices (and don't have my usual pack of doc with me).  A couple  
of notes:
  - external loopback is normally a diagnostic mode, so I don't know  
if the device would work differently than in "normal mode (does it  
require a loopback connector?); and
  - this definitely won't get you N-way (autonegotiation) support,  
but most (some?) switches are configurable for this feature (finding  
out about it is another trick).  However, hubs and switches won't  
default to FDX, so for general use, you're looking at some trouble.   
If you're using this in a limited, controlled environment, it may  
work.

Regards,

Justin

--
Justin C. Walker, Curmudgeon-At-Large *
Institute for General Semantics       |
Manager, CoreOS Networking            |   Men are from Earth.
Apple Computer, Inc.                  |   Women are from Earth.
2 Infinite Loop                       |       Deal with it.
Cupertino, CA 95014                   |
*-------------------------------------*-------------------------------*