Path: utzoo!news-server.csri.toronto.edu!rutgers!usc!zaphod.mps.ohio-state.edu!think.com!snorkelwacker.mit.edu!ira.uka.de!fauern!faui43.informatik.uni-erlangen.de!csbrod
From: csbrod@immd4.informatik.uni-erlangen.de (Claus Brod)
Newsgroups: comp.sys.atari.st.tech
Subject: Re: Bad.sector.list on HD-s?
Keywords: Bad sectors
Message-ID: <1991Mar11.130920.20617@informatik.uni-erlangen.de>
Date: 11 Mar 91 13:09:20 GMT
References: <2899@hexagon.se> <1991Mar8.003618.25646@jato.jpl.nasa.gov> <13466@ccncsu.ColoState.EDU> <1991Mar10.235310.8813@jato.jpl.nasa.gov>
Organization: CSD., University of Erlangen, Germany
Lines: 28

vsnyder@jato.jpl.nasa.gov (Van Snyder) writes:
>I've got software that lets me mark FAT entries to indicate bad clusters, and
>software that tries to discover which sectors are bad.  But the disk
>manufacturer uses a more stringent test than one can do with I/O through
>the SCSI.  I have an Adaptec 4000, and manual.  It's DEFINITELY possible to
>tell the controller about the bad spots.  The problem is, none of the *low-
>level* formatting programs I've tried know how to tell the controller about
>the manufacturer's defect list.  I'd much prefer to tell the Adaptec where the
>bad spots are in advance, than discover there's a bad spot in something I've
>just spent 20 hours creating.  (I do backups, but not continuously).  It's the
>*low-level* formatting program that's got to tell the Adaptec where the
>manufacturer says the bad spots are.  Unless somebody can provide a fool-proof
>(i.e. works with EVERYBODY's HD SW) algorithm for converting the "cylinder/
>head/bytes-from-index" indication of bad spots into partition/cluster-number
>indication.  BTW, marking FATs won't work the same way on Magic, GCR, Minix,
>etc. partitions, so telling the Adaptec seems the only reliable way.

I've been working on something like this, but I ended up concluding that
the Adaptec 4000 has firmware bugs in it that make it impossible to transfer\
a defect list when formatting. Recently, I received indications that there
might be a way to do it, but I have no details on that. Somebody out there
who can tell us the tricks?

----------------------------------------------------------------------
Claus Brod, Am Felsenkeller 2,			Things. Take. Time.
D-8772 Marktheidenfeld, West Germany		(Piet Hein)
csbrod@medusa.informatik.uni-erlangen.de
----------------------------------------------------------------------