From owner-cvs-sys  Wed May 13 06:29:25 1998
Return-Path: <owner-cvs-sys>
Received: (from daemon@localhost)
          by hub.freebsd.org (8.8.8/8.8.8) id GAA28100
          for cvs-sys-outgoing; Wed, 13 May 1998 06:29:25 -0700 (PDT)
          (envelope-from owner-cvs-sys)
Received: from godzilla.zeta.org.au (mothra.kralizec.net.au [203.15.68.22] (may be forged))
          by hub.freebsd.org (8.8.8/8.8.8) with ESMTP id GAA28035;
          Wed, 13 May 1998 06:28:50 -0700 (PDT)
          (envelope-from bde@godzilla.zeta.org.au)
Received: (from bde@localhost)
	by godzilla.zeta.org.au (8.8.7/8.8.7) id XAA06754;
	Wed, 13 May 1998 23:25:21 +1000
Date: Wed, 13 May 1998 23:25:21 +1000
From: Bruce Evans <bde@zeta.org.au>
Message-Id: <199805131325.XAA06754@godzilla.zeta.org.au>
To: bde@FreeBSD.ORG, bde@zeta.org.au, cvs-all@FreeBSD.ORG,
        cvs-committers@FreeBSD.ORG, cvs-sys@FreeBSD.ORG,
        luoqi@watermarkgroup.com
Subject: Re:  cvs commit: src/sys/i386/i386 autoconf.c src/sys/kern kern_shutdown.c
Sender: owner-cvs-sys@FreeBSD.ORG
X-Loop: FreeBSD.org
Precedence: bulk

>I guess we could make SLICE code do that. But eventually we'd like to get
>rid of dev_t all together, we have to find another way to figure out if
>a disk partition is a swap partition. IMO, using disklabel is the correct
>way (being a 'b' partition doesn't prevent user from creating a fs in it,
>and being a non-'b' partition doesn't mean you can't use it as a swap).
>It's just bad luck we couldn't retrieve disklabel during a dump.

I don't want to use a disk label.  I use a DOS partition (shared with
Linux) for swap and dumps.  I had to label it to get it to work under
FreeBSD (except under SLICEd kernel which don't recognise labels on Linux
swap partitions).  The label sector is essentially read-only in Linux -
the zero bits in it are interpeted as bad sectors, so I lose a little
swap space under Linux :-).

>I have a suggestion: since swap and dump code are the only consumers of
>d_psize(), can we change d_psize to return disklabel instead (similar to
>what DIOCGPART does)? This would also allow us to get rid of the DIOCGPART
>kludge (ioctl should only be called from userland). If you think this is
>doable, I can work out a patch for that.

No, raw disks don't have labels, and labels are too large.  DIOCGPART
and d_maxio should be replaced by additional interfaces like d_psize()
instead.  Eventually the interfaces should be something like:

#define	d_psize(diskdevp)	((diskdevp)->d_psize)

Bruce