From owner-freebsd-current  Fri May  3 15:11:25 1996
Return-Path: owner-current
Received: (from root@localhost)
          by freefall.freebsd.org (8.7.3/8.7.3) id PAA15484
          for current-outgoing; Fri, 3 May 1996 15:11:25 -0700 (PDT)
Received: from godzilla.zeta.org.au (godzilla.zeta.org.au [203.2.228.19])
          by freefall.freebsd.org (8.7.3/8.7.3) with SMTP id PAA15475
          for <current@FreeBSD.ORG>; Fri, 3 May 1996 15:11:20 -0700 (PDT)
Received: (from bde@localhost) by godzilla.zeta.org.au (8.6.12/8.6.9) id IAA19156; Sat, 4 May 1996 08:04:39 +1000
Date: Sat, 4 May 1996 08:04:39 +1000
From: Bruce Evans <bde@zeta.org.au>
Message-Id: <199605032204.IAA19156@godzilla.zeta.org.au>
To: asami@CS.Berkeley.EDU, rgrimes@GndRsh.aac.dev.com
Subject: Re: ccd offset, please review + test
Cc: ccd@stampede.cs.berkeley.edu, current@FreeBSD.ORG
Sender: owner-current@FreeBSD.ORG
X-Loop: FreeBSD.org
Precedence: bulk

> * And /dev/sd237c should _not_ be used.  You should manually go add a
> * /dev/sd237a and use that for the ccd, this should eliminate your problem.

>For the whole slice, or with an offset?  In other words, is sdXc
>special because of its name, or because it is the only partition that
>starts at the beginning?

sdXc is conventionally the whole slice.  Making it smaller than the whole
slice is fairly harmless and fairly useless.

>Wait, the latter doesn't make sense, all the machines here have the
>root filesystem starting at offset 0 (within the slice).  So you're
>saying sdXc is special because it has the letter `c' in it?

Grep for RAW_PART in /sys to see where sdXc is special.  It's special
before a label exists (then it's the only partition than can be opened)
and after a label exists (then the label on it is write protected ...).

> * Conventient, but wrong to do.  UNIX has reserved xxYc for as long as
> * I can remeber, using it for file systems is a sure fire way to burn
> * yourself.

>Well I don't think that is true, the SunOS machines I was

I think he means BSD.

>Anyway, I just tried creating a regular filesystem on /dev/sd237c
>(actually sd1c, but who's counting) on our FreeBSD machine.  It seems
>to work, are you sure it isn't supposed to?

I use /dev/rsds1 partitioned normally and /dev/rsds[2-4] essentially
unpartitioned (I have to put a label on them to keep newfs happy but
mount handles arbitrary block devices with a supported file system
on them).

Bruce