From owner-freebsd-mobile  Sat Oct 25 00:39:56 1997
Return-Path: <owner-freebsd-mobile>
Received: (from root@localhost)
          by hub.freebsd.org (8.8.7/8.8.7) id AAA19901
          for mobile-outgoing; Sat, 25 Oct 1997 00:39:56 -0700 (PDT)
          (envelope-from owner-freebsd-mobile)
Received: from word.smith.net.au (ppp20.portal.net.au [202.12.71.120])
          by hub.freebsd.org (8.8.7/8.8.7) with ESMTP id AAA19888
          for <mobile@freebsd.org>; Sat, 25 Oct 1997 00:39:42 -0700 (PDT)
          (envelope-from mike@word.smith.net.au)
Received: from word.smith.net.au (localhost [127.0.0.1])
	by word.smith.net.au (8.8.7/8.8.5) with ESMTP id RAA00394;
	Sat, 25 Oct 1997 17:05:50 +0930 (CST)
Message-Id: <199710250735.RAA00394@word.smith.net.au>
X-Mailer: exmh version 2.0zeta 7/24/97
To: Nate Williams <nate@mt.sri.com>
cc: Mike Smith <mike@smith.net.au>, mobile@freebsd.org
Subject: Re: Patches from -current for -stable I'd like to commit after testing 
In-reply-to: Your message of "Fri, 24 Oct 1997 10:43:42 CST."
             <199710241643.KAA20805@rocky.mt.sri.com> 
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Date: Sat, 25 Oct 1997 17:05:46 +0930
From: Mike Smith <mike@smith.net.au>
Sender: owner-freebsd-mobile@freebsd.org
X-Loop: FreeBSD.org
Precedence: bulk

> > ie. the mini-probe 
> > is basically going to run the probe and then attach routines again?
> 
> Just the probe, not the attach.

Then my basic gripe remains; in the ethernet case, if I pull card A and 
replace with card B of the same type, the arp code will be confused 
(wrong MAC address).

> > (Do you have to do this anyway, to get the PCCARD back to a known 
> > state?)
> 
> Well, there's the issue, and the answer is 'maybe so, I'm not sure'.  I
> don't *think* so, but it may require it.  I'm playing with some code to
> try and not require it.  I know the linux code doesn't try to save the
> state, and instead does what the apm_pccard_resume code does.  (Which
> isn't necessarily a bad thing.)  However, I'm not sure what the other
> OS's do (NetBSD for example).  Win95 appears to 'suspend/resume' the
> card, although it may just be the 'appearances', and not how it's
> actually implemented under the hood.

I think that caching "what was in the slot" at a lower level would be 
good, but that means moving the CIS parser inside the kernel.

mike