From owner-freebsd-bugs Sun Sep 7 06:33:53 1997 Return-Path: Received: (from root@localhost) by hub.freebsd.org (8.8.7/8.8.7) id GAA18148 for bugs-outgoing; Sun, 7 Sep 1997 06:33:53 -0700 (PDT) Received: from godzilla.zeta.org.au (godzilla.zeta.org.au [203.2.228.19]) by hub.freebsd.org (8.8.7/8.8.7) with ESMTP id GAA18143 for ; Sun, 7 Sep 1997 06:33:48 -0700 (PDT) Received: (from bde@localhost) by godzilla.zeta.org.au (8.8.5/8.6.9) id XAA06556; Sun, 7 Sep 1997 23:32:15 +1000 Date: Sun, 7 Sep 1997 23:32:15 +1000 From: Bruce Evans Message-Id: <199709071332.XAA06556@godzilla.zeta.org.au> To: bde@zeta.org.au, j@uriah.heep.sax.de Subject: Re: conf/4461: IOMega Zip disk entry for disktab Cc: freebsd-bugs@hub.freebsd.org Sender: owner-freebsd-bugs@FreeBSD.ORG X-Loop: FreeBSD.org Precedence: bulk >> > What's wrong with ``disklabel -Brw sdX auto''? I figure it should >> >> 1. `auto' shouldn't exist. Use >> `disklabel /dev/rsdX >foo; $EDITOR foo; disklabel -BrR sdX foo' >> instead. > >That's too hard unless you put it into a script that ships by default. >disklabel ... auto doesn't do much more, actually. It has a precedent >case in DEC Unix. Editors too hard? How about pipelines? The auto case is equivalent to the trivial pipleline `disklable /dev/rsdX | disklabel -BrR sdx /dev/stdin'. >> 2. It has the same problems as the committed disktab entry: >> a) it only works for "Dangerously Dedicated" disks. Note that the > >I don't think so. disklabel ... auto might even work for sliced disks >(but i'm not much experienced with sliced disks myself). I am experienced with sliced disks :-). For a factory-formatted zip disk with the type of the slice changed from 6 to 165, `disklabel -Brw sd1 auto' gives the bogus error message "write: Read-only file system", and the system logs the messages "fixlabel: raw partition size > slice size" and "fixlabel: raw partition offset != slice offset". For a factory- formatted zip disk with the type of the slice NOT changed from 6, `disklabel -Brw sd1 auto' works due to bugs (the DIOCSDINFO ioctl gives a wrong in-core label for /dev/rsd1c (one covering the whole disk although /dev/rsd1c has size 0) and the MBR gets clobbered :-(. I just found this bug. Bruce