From owner-freebsd-mobile@FreeBSD.ORG  Fri Dec 10 16:34:30 2004
Return-Path: <owner-freebsd-mobile@FreeBSD.ORG>
Delivered-To: freebsd-mobile@freebsd.org
Received: from mx1.FreeBSD.org (mx1.freebsd.org [216.136.204.125])
	by hub.freebsd.org (Postfix) with ESMTP id 1357816A4CE
	for <freebsd-mobile@freebsd.org>;
	Fri, 10 Dec 2004 16:34:30 +0000 (GMT)
Received: from smtp-sv.canvas.ne.jp (smtp-sv.canvas.ne.jp [210.172.56.252])
	by mx1.FreeBSD.org (Postfix) with ESMTP id 739A743D1D
	for <freebsd-mobile@freebsd.org>;
	Fri, 10 Dec 2004 16:34:29 +0000 (GMT)
	(envelope-from colinp@mx3.canvas.ne.jp)
Received: from canvas.ne.jp (unknown [192.168.100.61])
	by smtp-2.canvas.ne.jp (Postfix) with SMTP id E6CA015EAD8
	for <freebsd-mobile@freebsd.org>;
	Sat, 11 Dec 2004 01:34:27 +0900 (JST)
Received: (qmail 14221 invoked from network); 11 Dec 2004 01:34:27 +0900
Received: from unknown (HELO mail.mx3v.canvas.ne.jp) (192.168.100.53)
  by vc1d with SMTP; 11 Dec 2004 01:34:27 +0900
Received: from [127.0.0.1] (unknown [202.71.91.13])
	by mail.mx3v.canvas.ne.jp (Postfix) with ESMTP id 8A0B42C2B
	for <freebsd-mobile@freebsd.org>;
	Sat, 11 Dec 2004 01:34:27 +0900 (JST)
Message-ID: <41B9CFDC.3040305@mx3.canvas.ne.jp>
Date: Sat, 11 Dec 2004 01:33:32 +0900
From: Colin Peters <colinp@mx3.canvas.ne.jp>
User-Agent: Mozilla Thunderbird 0.8 (Windows/20040913)
X-Accept-Language: ja, en-us, en
MIME-Version: 1.0
To: freebsd-mobile@freebsd.org
References: <20041209120101.DF76716A4E3@hub.freebsd.org>
In-Reply-To: <20041209120101.DF76716A4E3@hub.freebsd.org>
Content-Type: text/plain; charset=ISO-8859-1; format=flowed
Content-Transfer-Encoding: 7bit
Subject: Re: corega FEtherII PCC-TXD
X-BeenThere: freebsd-mobile@freebsd.org
X-Mailman-Version: 2.1.1
Precedence: list
List-Id: Mobile computing with FreeBSD <freebsd-mobile.freebsd.org>
List-Unsubscribe: <http://lists.freebsd.org/mailman/listinfo/freebsd-mobile>,
	<mailto:freebsd-mobile-request@freebsd.org?subject=unsubscribe>
List-Archive: <http://lists.freebsd.org/pipermail/freebsd-mobile>
List-Post: <mailto:freebsd-mobile@freebsd.org>
List-Help: <mailto:freebsd-mobile-request@freebsd.org?subject=help>
List-Subscribe: <http://lists.freebsd.org/mailman/listinfo/freebsd-mobile>,
	<mailto:freebsd-mobile-request@freebsd.org?subject=subscribe>
X-List-Received-Date: Fri, 10 Dec 2004 16:34:30 -0000

A few more bits of data from my continuing experiments.

The extra write of 0x04 to 0x3c2 (shown below) appears to be necessary 
for the FEtherII PCC-TXD. Without it the card refuses to respond (the 
link light stays off and I get "ed1: device timeout" messages again).

The other thing I have found is that ed_pccard_match is getting called 
and setting the flags it wants to set, but those flags are cleared 
before getting to ed_pccard_probe.

static int
ed_pccard_match(device_t dev)
{
 ...
                if (pp->flags & NE2000DVF_DL10019)
                        device_set_flags(dev, ED_FLAGS_LINKSYS);
                else if (pp->flags & NE2000DVF_AX88190)
                        device_set_flags(dev, ED_FLAGS_AX88190);
printf(" in ed_pccard_match flags=%08x\n", device_get_flags(dev));
                return (0);


Results in this in dmesg:

 in ed_pccard_match flags=00030000
ed1: <corega K.K. corega FEtherII PCC-TXD> at port 0x300-0x31f irq 9 
function 0 config 7 on pccard1


But this code in probe shows the flags are no longer set at that point, 
which is why I had to force ed_pccard_ax88190 to be called manually.

static int
ed_pccard_probe(device_t dev)
{
        int     error;
        int     flags = device_get_flags(dev);

device_printf(dev, "pccard probe");
printf(" flags=%08x compare %08x.\n", flags, ED_FLAGS_AX88190);
/*      if (ED_FLAGS_GETTYPE(flags) == ED_FLAGS_AX88190) { */
                error = ed_pccard_ax88190(dev);
                goto end2;
/*      } */


The output of dmesg shows:

ed1: pccard probe flags=00000000 compare 00030000.


Now I will try to figure out where these functions are called and why 
the flags are cleared between them. Perhaps it would be better to call 
match again at the beginning of probe (or use a similar approach to 
determine the device type at probe time)?

Colin.