Return to BSD News archive
Path: sserve!newshost.anu.edu.au!munnari.oz.au!metro!sequoia!ultima!kralizec.zeta.org.au!kralizec.zeta.org.au!not-for-mail
From: bde@kralizec.zeta.org.au (Bruce Evans)
Newsgroups: comp.os.386bsd.questions
Subject: Re: Problems with NetBSD-08 and ESDI - Please help!
Date: 17 Aug 1993 05:55:37 +1000
Organization: Kralizec Dialup Unix Sydney: +61-2-837-1183 V.32bis
Lines: 65
Message-ID: <24oonpINNffi@kralizec.zeta.org.au>
References: <244ule$cke@netbsd08.telecom.com.au> <5d3J03pYd8aM00@amdahl.uts.amdahl.com>
NNTP-Posting-Host: kralizec.zeta.org.au
In <5d3J03pYd8aM00@amdahl.uts.amdahl.com> agc@uts.amdahl.com (Alistair G. Crooks) writes:
>In article <244ule$cke@netbsd08.telecom.com.au> tdwyer@netbsd08.telecom.com.au (Terry Dwyer 491 5161) writes:
>>
>>After having installed NetBSD-08 on two 486's with Adaptec 1542B SCSI
>>controllers, I thought I would not have too many problems installing
>>on a 486 with a Western Digital WD1007V ESDI controller. Boy, was I
>>wrong. The drive is a WREN V 94186-383H. Translation has been turned
>>off. I have tried alternate sector mapping on and off, set the drive SPT
>>(according to manufacturers specs) to 34 and 36, used the ESDI bios to
>>low level format the drive, and WDFMT (a WD utility), to do the same.
I use a similar drive/controller. I have alternate sector mapping turned
off (why waste a sector or two to get the controller's braindamaged sector
forwarding when you can waste far less to get the bad144 braindamaged
sector forwarding? :-). I have translation enabled but use the identity
translation except for testing translation.
>>... It seems that NetBSD
>>is not paying any attention to any bad sector mapping done by either of
>>the WD formatters. When installing NetBSD, I've tried a y and an n reply
>>to the question about automatic bad sector forwarding, results were the same
>>both times, with bad sectors being found either during install or when copying
>>the distribution onto the drive.
Perhaps NetBSD just sets the sector formatting flag but doesn't actually
put the bad sectors in the bad144 table. It would need to read the entire
disk to do that, and you would probably notice the delay.
>You seem to be having the same problem that I've got with an "IDE"
>drive that refuses to forward bad sectors automatically. Basically,
>you've got a disc with 1747 (I think) cylinders, but the BIOS puts an
>upper limit of 1024 sectors on the number of cylinders you can have.
>[This causes problems booting because the bootstrap uses the BIOS and
> the BIOS doesn't know about cylinders >= 1024.]
I think that's a different problem.
>The best way to get NetBSD 0.8 on your system would be to tell the
>install program that you've a disc of 1024 cylinders. (That should let
>you use about 65% of the total space on the disc).
Why waste 41.4% of the disk? The following hack might work: lie about the
siae of the 'c' partition in the disk label. Make it end on cylinder 1023
or lower. Keep the working partitions away from the the bad144 table and
spare sectors at the end of partition 'c' (reserving one cylinder usually
works). E.g.:
BIOS partition 1: another OS 1: cylinders 0-99
BIOS partition 2: another OS 1: cylinders 100-199
BIOS partition 3: another OS 1: cylinders 200-299
BIOS partition 4: 386BSD: cylinders 300-1746: a: 300-399 (boot)
b: 400-499 (swap)
c: 300-1023 ("386BSD")
d: 0-1746 (all)
e: 1023-1023 (bad144)
f: 500-1022 (usr)
g: 1024-1746 (local)
This is completely untested.
Bruce
--
Bruce Evans bde@kralizec.zeta.org.au