From owner-freebsd-hackers  Wed Jul 10 14:52:02 1996
Return-Path: owner-hackers
Received: (from root@localhost)
          by freefall.freebsd.org (8.7.5/8.7.3) id OAA18309
          for hackers-outgoing; Wed, 10 Jul 1996 14:52:02 -0700 (PDT)
Received: from godzilla.zeta.org.au (godzilla.zeta.org.au [203.2.228.19])
          by freefall.freebsd.org (8.7.5/8.7.3) with SMTP id OAA18277
          for <hackers@freebsd.org>; Wed, 10 Jul 1996 14:51:52 -0700 (PDT)
Received: (from bde@localhost) by godzilla.zeta.org.au (8.6.12/8.6.9) id HAA28066; Thu, 11 Jul 1996 07:46:32 +1000
Date: Thu, 11 Jul 1996 07:46:32 +1000
From: Bruce Evans <bde@zeta.org.au>
Message-Id: <199607102146.HAA28066@godzilla.zeta.org.au>
To: SimsS@Infi.Net, davidg@Root.COM
Subject: Re: Some recent changes to GENERIC
Cc: hackers@freebsd.org
Sender: owner-hackers@freebsd.org
X-Loop: FreeBSD.org
Precedence: bulk

>>(For my benefit, can anyone explain why the default port address for ed0
>>is 0x280?  I assume some historical reason, but I really don't know the
>>details....)

Perhaps it was to avoid a conflict with another (non-ed) ethercard at
0x300, but nothing is gained if ed1 is on 0x300.

>>OK, how 'bout this:  I'll endorse nuking the ed1 device, but only if the
>>default settings for the ed0 device are changed to 5/300 in GENERIC.  What
>>say?

>   This seems reasonable to me. 0x300 is supported by all of the ed-compatible
>cards as far as I know.

I use 0x300/15 because interrupts below 8 are too scarce towaste on 16-bit
cards, so GENERIC kernels never work without boot time configuration here.
The factory default is 0x300/3 (irq 3 is the default because the board is
designed to work in 8-bit slots and irq 5 is not available on XT's).  The
wrong GENERIC ed0 works better than the wrong GENERIC ed1 here because
the failure is more obvious :-) - when I forget to change the irq, the
ed1 probe works because the address is right and interrupt timeouts occur
later.

Bruce