Path: utzoo!utgpu!jarvis.csri.toronto.edu!mailrus!ncar!tank!uwvax!per2!dag
From: dag@per2.UUCP (Daniel A. Glasser)
Newsgroups: comp.sys.atari.st
Subject: Re: 2K of disk space lost?
Summary: A possible safe way around the lost clusters...
Message-ID: <837@per2.UUCP>
Date: 4 Apr 89 14:58:54 GMT
References: <8903211617.AA27122@uk.ac.ucl.sm.uxm> <54471@yale-celray.yale.UUCP>
Organization: Persoft Inc., Madison, WI
Lines: 41

In article <54471@yale-celray.yale.UUCP>, root@yale.UUCP (Celray Stalk) writes:
> In article <8903211617.AA27122@uk.ac.ucl.sm.uxm> andrew@uxm.sm.ucl.ac.UK (Andrew Dawson)
> writes:
> >Is this a TOS bug, or have I missed something? Anyone got any ideas?
> 
> Yes, this is a known TOS bug.  The last two clusters of every disk
> remain unused. [...]
> [...]  Thus, last I heard, Atari had decided *not* to fix the
> problem in TOS 1.4.
> ==================================================
> | Michael Fischer                                |
> ==================================================

Back in mid-1986, I wrote a RAM-disk utility at Mark Williams Co., which
became the basis for the RDY utility (I did not write RDY, that was done
by Roger, who is still working for MWC).  I was aware of this bug in the
TOS cluster allocation scheme, and built a fix into the FAT initialization
code in the non-resident part of the program.  This involved telling the
OS that there were two extra clusters (in the disk configuration in block 0)
and marking the last two clusters as bad in the FAT.  The OS would never
try and access these clusters, since they were "known" to be bad, and the
buggy OS code would have no trouble allocating the last two real clusters
because the bug is in the determination of which FAT entry corresponds to
which cluster on allocation only.  I don't remember if I ever tried this
on a floppy, I know I never tried it on a hard drive.  It is my belief,
however, that the kluge will work and won't break on a fixed OS.
The source for RDY includes this hack as an option.  If you have the MWC
package, you may want to look into the source.  If you want more details,
send me mail.  I don't have my ST here at my office, so I can't give any
real details right now.

I hope this helps someone.  Alan -- What's your opinion of this hack?
Any problems that I missed?  Also, a question about TOS 1.4 -- Can
it handle cluster sizes != 2 sectors?

						Daniel A. Glasser
-- 
 _____________________________________________________________________________
    Daniel A. Glasser                           One of those things that goes
    uwvax!per2!dag                              "BUMP!!!(ouch)" in the night. 
 ---Persoft, Inc.---------465 Science Drive-------Madison, WI 53711-----------