Date: Wed, 29 Nov 1995 13:55:36 -0700 (MST) From: Terry Lambert <terry@lambert.org> To: wwong@wiley.csusb.edu (William Wong) Cc: questions@FreeBSD.ORG Subject: Re: reposting again... Message-ID: <199511292055.NAA28570@phaeton.artisoft.com> In-Reply-To: <199511291039.CAA10531@wiley.csusb.edu> from "William Wong" at Nov 29, 95 02:39:13 am
next in thread | previous in thread | raw e-mail | index | archive | help
> I'm trying to get the DPT ESDI caching controller (PM3011) w/4 Megs working > with 2 Fujitsu 670 Meg Full Height (fast ESDI) drives. Well, at least that is > what the person who owns this ancient beast told me, that is, the fast ESDI > part. After low level formatting the drives with the supplied utilities from > DPT and booting up from the FreeBSD boot disk, the drives are seen as with an > "unknown size, using BIOS defaults". No matter how many cylinders I specify in > the BIOS, FreeBSD's boot disk only sees 1025 cylinders. I guess the machine > just has an old BIOS. Anyway, I decided to go with 1025 cylinders just to see > if the drive can be partitioned and newfs'd. I keep getting errors with > "command return status of 36". I'm using 1025/15/53 for C/H/S since 1652/15/53 > isn't being seen. Are these the correct values for these Fujitsu drives? Are > there any special settings I need to set for the DPT controller? I don't know > if BIOS translation and sector sparing can be disabled on this controller. I > am using the whole drive for FreeBSD. Actually, both of them. > > Comments, anyone? This is the same error I got on a "seek to sector past end of track" on a wD1007 ESDI controller with sector sparing turned on. The problem is that the physical geometry reported by the controller "get drive geometry" includes the "spared" sectors. Generally, you will have to turn off sector sparing and geometry translation for the drive for it to work. This is true for most older drives that use BIOS-based sector sparing and/or geometry translation mechanisms. The number of cylinders you will need to tell it is 1023. This is to make the fdisk happy. You can later fix this in the disklabel (they are unfortunately "integrated" in the new install). Typically, this means making a seperate /usr, and expanding its size and newfs'ing by hand using the shell on the alternate console later. Or you can modify the disklabel after writing it in the install tools. To know the actual number of cylinders, you'll need to boot with the -v flag or run pfdisk.exe from DOS (available from the same place you got your FreeBSD boot disk). If you get a bad sector in the MBR or partition table, you will be screwed. If you get a bad sector in the second stage boot or the disklabel, you're screwed again. If you get a bad sector in slice 'a', the sector will be replaced, but if your FreeBSD area goes past cylinder 1023, the spare sector will be out of range of BIOS, and you won't be able to boot. These problems are a consequence of the sector sparing being done at the slice rather than the FreeBSD area level, and the bad sector spares being located at the end of the disk instead of at the end of slice 'a'. Hopefully this will be corrected some day. Terry Lambert terry@lambert.org --- Any opinions in this posting are my own and not those of my present or previous employers.
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?199511292055.NAA28570>