From owner-freebsd-current  Wed May 31 21:59:44 1995
Return-Path: current-owner
Received: (from majordom@localhost)
          by freefall.cdrom.com (8.6.10/8.6.6) id VAA28648
          for current-outgoing; Wed, 31 May 1995 21:59:44 -0700
Received: from gndrsh.aac.dev.com (gndrsh.aac.dev.com [198.145.92.241])
          by freefall.cdrom.com (8.6.10/8.6.6) with ESMTP id VAA28640
          for <current@FreeBSD.org>; Wed, 31 May 1995 21:59:42 -0700
Received: (from rgrimes@localhost) by gndrsh.aac.dev.com (8.6.11/8.6.9) id VAA03735; Wed, 31 May 1995 21:58:16 -0700
From: "Rodney W. Grimes" <rgrimes@gndrsh.aac.dev.com>
Message-Id: <199506010458.VAA03735@gndrsh.aac.dev.com>
Subject: Re: -cur boot hangs since may 1
To: bde@zeta.org.au (Bruce Evans)
Date: Wed, 31 May 1995 21:58:16 +1700 (PDT)
Cc: bugs@ns1.win.net, current@FreeBSD.ORG
In-Reply-To: <199505172031.GAA28969@godzilla.zeta.org.au> from "Bruce Evans" at May 18, 95 06:31:33 am
X-Mailer: ELM [version 2.4 PL24]
MIME-Version: 1.0
Content-Type: text/plain; charset=US-ASCII
Content-Transfer-Encoding: 7bit
Content-Length: 1375      
Sender: current-owner@FreeBSD.ORG
Precedence: bulk

> 
> >If you disable ie0 during a '-c' boot dialog (or build a kernel without ie0) 
> >then things work great.
> 
> >All my boxes have ep0 cards so I have been an unhappy boy for many days.
> 
> >I am looking at the probe technique in the ie0 driver to see what has
> >been changed.  It may be confusing the 3c509 card.
> 
> This is a known bug and was worked around in revision 1.20 (1994/11/18)
> of GENERIC.
> 
> 	Log:
> 	Put ie0 above ep0.  Otherwise, the ie0 probe clobbers it.
> 
> The work around was broken by sorting the devices in revision 1.38
> (1995/04/08).

Thanks for the leg work Bruce, I have commited what I think should fix
this and document it well enough that it does not happen until some
one fixes these probes up to not smash each other.

> The conflict is probably because      if_ie.c uses port ELINK_ID_PORT which
> is defined in elink.h    as 0x100 and if_ep.c uses port    EP_ID_PORT which
> is defined in if_epreg.h as 0x100.  The conflict is not detected because
> use of these ports is not recorded in the config file (there is no room
> for it) and the base ports don't conflict.  The conflict is not obvious
> in isa.h because few or no network ports are recorded there.
> 
> Bruce
> 


-- 
Rod Grimes                                      rgrimes@gndrsh.aac.dev.com
Accurate Automation Company                   Custom computers for FreeBSD