1996-01-27 - Re: Wipe Swap File

Header Data

From: dlv@bwalk.dm.com (Dr. Dimitri Vulis)
To: cypherpunks@toad.com
Message Hash: a9ee33d8ecd73ffc159765efbf510ca71aaac8b2e3aac6734d1f0e99e4b536ae
Message ID: <FucciD58w165w@bwalk.dm.com>
Reply To: <m0teD3N-00090UC@pacifier.com>
UTC Datetime: 1996-01-27 04:37:59 UTC
Raw Date: Sat, 27 Jan 1996 12:37:59 +0800

Raw message

From: dlv@bwalk.dm.com (Dr. Dimitri Vulis)
Date: Sat, 27 Jan 1996 12:37:59 +0800
To: cypherpunks@toad.com
Subject: Re: Wipe Swap File
In-Reply-To: <m0teD3N-00090UC@pacifier.com>
Message-ID: <FucciD58w165w@bwalk.dm.com>
MIME-Version: 1.0
Content-Type: text/plain


(This came in on Sunday and I mislaid it, sorry)

jim bell <jimbell@pacifier.com> writes:
> >Jim Bell mentioned the trick of hiding information into 'extra' tracks and
> >sectors not used by the usual DOS formatting. It's very old too.
>
> Actually, I think it wasn't really used initially for "data hiding"
> purposes.  I'm talking about the early days of CP/M and other such systems,
> circa 1977 and such, when individuals "discovered" that floppy drives had no
> hard mechanical stop past the "last" good track, and they "stole" a few
> percent of extra capacity from a floppy by simply ignoring the recommended
> "last" track.  Naturally, it would work okay on some drives but not on
> others... which is why it was a bad idea.

All the 360K 5.25" drives I've ever tried were capable of going to track 43
reliably (regular PC DOS format used 40). But I've come across too many
1.44/3.5" drvies that don't go beyond 80. Can't count on it. However...

> In addition, I also discovered that it was possible to put a few more than
> 26 sectors on each track of an 8" single-density (240 kilobytes!) floppy
> disk.  The main problem with using
> these "tricks" is that the floppy had no method of conveying formatting
> information to the system it was in, which meant that any floppy using this
> trick was by definition non-standard.  ("feature" or "bug" depending on your
> goal...)

It's possible to squeeze more sectors on 5.25" and 3.5" disks too, especially
if you specify shorter gaps. It's also possible to vary sector size.

When I read the above paragraph, I thought momentarily: it takes A LOT of bytes
to specify the format of a floppy disk. Could one use this as a kind of
encryption? I.e., the key would be the format specs, and it would have to
be supplied to the device driver before it could read the actual data from
the floppy. Unfortunately, I think a clever analysis of the disk with just
the regular FDC can tell you a lot about the formatting (i.e., the number and
the size of the sectors, and even their physical order).

> I started building my own 12.5 MHz Z-80 -based CP/M system in 1978, fully
> designed and wire-wrapped by myself, and wrote my own BIOS.  (Used a WDC
> 1791 FDC)  Had total
> control.  I didn't try this trick even then because of compatibility
> reasons, but one thing I _DID_ do was to write a floppy formatter that
> "undid" the 6-sector skewing that standard CP/M had to do to keep up with
> the data read/write. (in other words, I physically re-skewed the sector
> numbering to make the next "desired" sector come faster...) I ended up with a
> effective skew-factor of 2.  Even a skew factor of 1 worked on my system (no
> skew at all), but the problem was that when I gave the most extreme of these
> oddly skewed floppies to my friends with 8" floppies, they took A LONG TIME
> to read the data!  (Their systems always missed the next sector because their
> systems were too slow, so they only ended up being able to read one sector
> per disk rotation.)

This is extremely cool indeed.

Trivia question: is it true that the reason why a lot of hard disks have
a prime number of sectors per track (like 17) so that you can use different
interleave factors (which have to be relatively prime to the # of sectors)?

> All this helps to explain why I asked if PGP had ever been ported to CP/M.
> Nostalgia!

Only a few weeks ago I gave away a NEC V20 motherboard capable of running CP/M.
(I recall running CP/M on it at some point. Of course, its main purpose was
being an 8088 clone.) Why not. :-)

It would be even cooler to port PGP to BESM-6. (How many people on this
list have ever used a BESM-6?)

---

Dr. Dimitri Vulis
Brighton Beach Boardwalk BBS, Forest Hills, N.Y.: +1-718-261-2013, 14.4Kbps





Thread