From owner-freebsd-mobile  Fri Apr  7 23: 9:22 2000
Delivered-To: freebsd-mobile@freebsd.org
Received: from rover.village.org (rover.village.org [204.144.255.49])
	by hub.freebsd.org (Postfix) with ESMTP
	id AC17F37BAB3; Fri,  7 Apr 2000 23:09:18 -0700 (PDT)
	(envelope-from imp@harmony.village.org)
Received: from harmony.village.org (harmony.village.org [10.0.0.6])
	by rover.village.org (8.9.3/8.9.3) with ESMTP id AAA09102;
	Sat, 8 Apr 2000 00:09:17 -0600 (MDT)
	(envelope-from imp@harmony.village.org)
Received: from harmony.village.org (localhost.village.org [127.0.0.1]) by harmony.village.org (8.9.3/8.8.3) with ESMTP id AAA09623; Sat, 8 Apr 2000 00:08:23 -0600 (MDT)
Message-Id: <200004080608.AAA09623@harmony.village.org>
To: mobile@freebsd.org
Cc: mdodd@freebsd.org
Subject: Two pccards in the tree for a while
Date: Sat, 08 Apr 2000 00:08:23 -0600
From: Warner Losh <imp@village.org>
Sender: owner-freebsd-mobile@FreeBSD.ORG
Precedence: bulk
X-Loop: FreeBSD.org


I've been thinking about driver attachments in the NEWCARD world.  I'd
like to make them as compatible as I can to the oldcard world.  To
that end, I'd like to propose a small addition to each of the pccard
driver attachments in anticipation of NEWCARD.  Right now the typical
pccard_probe routine looks like one of the following:

foo_pccard_probe(device_t dev)
{
	device_set_description(dev, "foo pccard thing");
	return 0;
}

		or

foo_pccard_probe(device_t dev)
{
	return foo_probe(dev);
}

Where foo_probe might be expanded inline foo_pccard_probe.

I'd like to propose a small addition.  I'd like to move to primarily
identifying the cards by their OEM number and OEM ID (these numbers
are similar to the pci vendor codes).  Each driver would have a table
of IDs that they support.  So the above driver's probe routines would
turn into

foo_pccard_probe(device_t dev)
{
	if (pccard_probe_vendor_id(dev, &foo_vendor_table))
		return ENXIO;
	/* Optionally probe the hardware here */
	device_set_description(dev, "foo card thang");
	return 0;
}

I'd also provide a pccard_probe_fucntion_id() as well for ata and sio
to use.  For existing pccard drivers, the changes are trivial to have
them continue to work with OLDCARD.  More work if you want to get a
good, accurate list of vendor ids, but NetBSD can be a good place to
look.

I may also provide a pccard_probe_cis_strings() routine, since I know
there are some cards that are only identifiable by their CIS strings,
but these are few and far between.

Of course, it is highly desirable to have the ability for a driver to
read its own CIS in various ways, and as that is accomplished, the
results will be available in both OLDCARD and NEWCARD.  The xe driver
is the only one right now that does this, but there are other drivers
that would like to do this, but they have pccardd provide them with
information at the moment for things like ethernet address.

However, in the old world order, pccardd desides what card gets what
drivers.  So, these new functions would be implemented as return 0; in
OLDCARD so that we can use the same drivers (in source) for both the
old and new code.  There may be other problems that crop up along the
way.  This is the only one that I'm seeing on my horizon right now, so
I thought I'd get people talking about these issues before I get to
the point of needing to jump on them.  I have the CIS reading code
approaching completion and I think the next thing after that is driver
attachment.  I'll know more in a few days.

I think the transition period in -current is going to be long enough
that these issues are important.  It would also be nice if we can use
the same or substantially drivers in -stable and -current for as long
as possible.

Comments?

Warner




To Unsubscribe: send mail to majordomo@FreeBSD.org
with "unsubscribe freebsd-mobile" in the body of the message