From owner-freebsd-scsi Mon Jan 26 09:57:06 1998 Return-Path: Received: (from majordom@localhost) by hub.freebsd.org (8.8.8/8.8.8) id JAA25450 for freebsd-scsi-outgoing; Mon, 26 Jan 1998 09:57:06 -0800 (PST) (envelope-from owner-freebsd-scsi@FreeBSD.ORG) Received: from narnia.plutotech.com (narnia.plutotech.com [206.168.67.130]) by hub.freebsd.org (8.8.8/8.8.8) with ESMTP id JAA25437 for ; Mon, 26 Jan 1998 09:57:02 -0800 (PST) (envelope-from gibbs@narnia.plutotech.com) Received: (from gibbs@localhost) by narnia.plutotech.com (8.8.8/8.7.3) id KAA27413; Mon, 26 Jan 1998 10:54:17 -0700 (MST) Date: Mon, 26 Jan 1998 10:54:17 -0700 (MST) From: "Justin T. Gibbs" Message-Id: <199801261754.KAA27413@narnia.plutotech.com> To: Bruce Evans cc: scsi@FreeBSD.ORG Subject: Re: Minor annoyances with ZIP Plus Newsgroups: pluto.freebsd.scsi In-Reply-To: <199801250804.TAA12845@godzilla.zeta.org.au> User-Agent: tin/pre-1.4-971204 (UNIX) (FreeBSD/3.0-CURRENT (i386)) Sender: owner-freebsd-scsi@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.org >>asking for the rigid geometry >>wasn't buying us anything and the code was recently changed to use the >>information returned by the Read Capacity command instead. > > This hasn't actually changed (at least in the current sd driver). All > that changed recently is that Read Capacity is now used to determine the > sector size. Yes. It's my fault for thinking in terms of what the CAM disk driver has been doing for some time. > Read Capacity has always been used to determine the disk size, and Mode > Sense of the rigid geometry page has always been used to initialize > default values for the geometry in the dummy label for the whole disk. > If there is a valid DOS partition on the disk, as is normally(?) the > case for new zip disks, then these default values are not used. In any > case, the geometry in the dummy label for fdisk is only used by fdisk > and sysinstall. Which is exactly the reason why we shouldn't be using the RG page at all. Almost every SCSI adapter we support will not properly boot a system with a physical geometry in the MBR and when you get a new disk drive it doesn't have a valid MBR to override the dummy label. This is why CAM offers a facility to determine the controller translated geometry and to use that in the dummy label. > Summary: everything just works, but the lowest layer is too verbose and > the highest layer (newfs) is too smart. If it just worked, we wouldn't get so many complaints about geometry problems during installs. > Bruce -- Justin