From owner-freebsd-current  Sat May 27 08:57:16 1995
Return-Path: current-owner
Received: (from majordom@localhost)
          by freefall.cdrom.com (8.6.10/8.6.6) id IAA24236
          for current-outgoing; Sat, 27 May 1995 08:57:16 -0700
Received: from gw.itfs.nsk.su (gw.itfs.nsk.su [193.124.36.33])
          by freefall.cdrom.com (8.6.10/8.6.6) with ESMTP id IAA24229
          for <current@freebsd.org>; Sat, 27 May 1995 08:57:04 -0700
Received: (nnd@localhost) by gw.itfs.nsk.su (8.6.12/8.6.12) id WAA07952 for current@freebsd.org; Sat, 27 May 1995 22:56:35 +0700
Date: Sat, 27 May 1995 22:56:35 +0700
From: "Nickolay N. Dudorov" <nnd@gw.itfs.nsk.su>
Message-Id: <199505271556.WAA07952@gw.itfs.nsk.su>
To: current@FreeBSD.org
Subject: ep and ie drivers conflict (?)
Sender: current-owner@FreeBSD.org
Precedence: bulk

	Yesterday I try to boot so called BOOTMFS
kernel (this is a GENERIC kernel with MFS_ROOT from
the current boot_floppy) on a computer with 3C509 card.

After probing (succesfully) all the devices the system says
"changing root to fd0c" and hangs. Booting with '-c' and
disabling various (absent in my computer) devices I found
that this kernel succesfully boots (and starts 'sysinstall')
if I disable 'ie0' device.

It seems to me that problem is in 3COM's 'ELINK_ID_PORT'
or 'EP_ID_PORT' (0x100) which is used in both ep and ie
drivers. (Is it right that probe order of devices follows
their order in config file ?).

This problem may be solved by:

 1) creating some common '3COM_probe' routine which'll be
    called by various drivers but performs manipulations
    with ELINK_ID_PORT only once, OR

 2) rewriting the probe parts of 'ep' and 'ie' drivers
    to leave the cards in some 'right' state after
    successfull AND unsuccessful probing
    (if it is possible at all), OR

 3) using some other method :-).

But I fear that all this'll be too late for 2.0.5.

Is there anybody to prove that I'm wrong ?

	N.Dudorov