[comp.sys.ibm.pc] cpio for DOS and 3b2?

samperi@unirot.UUCP (10/03/87)

I just finished implementing a UNIX-compatible cpio for DOS on a PC, and
I hoped to use it to transfer files via floppy disks between my AT
(running System V UNIX) and AT&T 3B2's (also running System V UNIX). But
I've since learned that the 3B2's do not format floppies in the same way
that the PC does. Does anyone know how floppies are formatted on the 3B2?
Is there software available that will enable a PC to format a floppy so
that a 3B2 can use it, or vis versa?

Dominick Samperi


UUCP: ...!ihnp4!cmcl2!manhat!samperi
      ...!rutgers!unirot!samperi

bhj@bhjat.UUCP (burt janz) (10/04/87)

In article <472@unirot.UUCP>, samperi@unirot.UUCP (Dominick Samperi) writes:
> I just finished implementing a UNIX-compatible cpio for DOS on a PC, and
> I hoped to use it to transfer files via floppy disks between my AT
> (running System V UNIX) and AT&T 3B2's (also running System V UNIX). But
> I've since learned that the 3B2's do not format floppies in the same way
> that the PC does. Does anyone know how floppies are formatted on the 3B2?

Recently, I tried to read some 3b2 floppies on Microport SV/AT V2.2.2L on
my AT clone.  I was able to read the first track...

cpio -icvt6

meaning that they use the old "version 6" type of CPIO format.  However,
with both the 360k floppy and the 1.2mb floppy I got read errors when the
floppy stepped to the next track.

I get the feeling that it is more than a floppy format problem.  I think
that they must be using a non-standard 360k floppy drive.

Admittedly, I did NOT investigate further, since I felt that swapping drives
and spending over 3 hours just trying to read 3 floppies was not worth the
effort to get the data on the floppies.

Hope this helps.

Burt Janz

jdu@ihopa.ATT.COM (John Unruh) (10/05/87)

The 3B2 does not use a 360K drive.  It uses a higher density drive
(you have to use 1.2 Mb floppies to make it work reliably).  I think
it holds about 760K on a disk.
                             John Unruh

wnp@killer.UUCP (Wolf Paul) (10/05/87)

In article <472@unirot.UUCP>, samperi@unirot.UUCP (Dominick Samperi) writes:
> I just finished implementing a UNIX-compatible cpio for DOS on a PC, and
> I hoped to use it to transfer files via floppy disks between my AT
> (running System V UNIX) and AT&T 3B2's (also running System V UNIX). But
> I've since learned that the 3B2's do not format floppies in the same way
> that the PC does. Does anyone know how floppies are formatted on the 3B2?
> Is there software available that will enable a PC to format a floppy so
> that a 3B2 can use it, or vis versa?

Two pieces of information on this:

1. Rich Salz recently posted Emmet Gray's MTOOLS package to comp.sources.unix.
This package will allow any UNIX system to read MS-DOS diskettes, provided
that the UNIX system has a floppy drive physically capable of reading the
MS-DOS formats (i.e. 40 Tracks @ 9 sectors for 360K disks). This would allow
you to copy a DOS file to the 3B2 or vice versa, and then unpack it with the
appropriate archiver.

2. The MKS Toolkit contains a dd(1) command, which permits reading and
writing floppy disks without reference to the file system on the disk,
i.e. it just reads/writes sectors in the proper sequence. With this command,
I can write an archive to a floppy from DOS like this:

  find ... | cpio ... | dd of=a:

and read the archive on the UNIX system with 

  cpio -icdm < /dev/rdsk/fd048      # if a 360K disk
or
  cpio -icdm < /dev/rdsk/fd096      # if a 1.2M disk

Again, this requires that the UNIX system have disk drives (and drivers)
capable of reading the raw MS-DOS format(s).

The formats are 40 Tracks @ 9 sectors for 360K, and 80 tracks @ 15 sectors
for 1.2MB. On Microport System V/AT, the RAW disk device entries are in
/dev/rdsk/fd*, and have names indicative of the drive characteristics
as follows:

    fd{dr}{dens}[{sides}{secs}]

where

   dr          = Drive #, either 0 for A: or 1 for B:
   dens        = Density in tracks/inch, 48 for double density, 96 for 
                 quad/high density. Note that these numbers translate
                 40 and 80 tracks respectively PER DISK, since a 5.25 floppy
                 does not use a full inch if tracks.
   sides       = ds for doublesided, ss for singlesided. All current DOS
                 formats (360K, 720K, 1.2M) are doublesided.
   secs        = Sectors per track, 9 for 360K and 720K disks, 15 for
                 1.2M disks.

I do not know how these naming conventions compare to those on a 3B2,
but I am sure that your documentation will specify the raw track/sector
format of the various floppies your system can read.

Wolf Paul
ihnp4!killer! [ mcomp! ] wnp

randy@chinet.UUCP (Randy Suess) (10/05/87)

In article <582@ihopa.ATT.COM> jdu@ihopa.ATT.COM (John Unruh) writes:
>The 3B2 does not use a 360K drive.  It uses a higher density drive
>(you have to use 1.2 Mb floppies to make it work reliably).  I think
>it holds about 760K on a disk.

	Microport can read/write 3b2 floppies and vice/versa.
	You need to create a new entry in /dev/dsk and /dev/rdsk
	It is something like minor 86 or 87.  I gave up on Microport
	months ago, so cannot give the exact number, but one of those
	are correct.  Microport tech support should also have
	the info.  Cpio -oc files are then easily transferable.

-- 
that's the biz, sweetheart.....
Randy Suess
..!ihnp4!chinet!randy

dave@westmark.UUCP (Dave Levenson) (10/06/87)

In article <151@bhjat.UUCP>, bhj@bhjat.UUCP (burt janz) writes:
> In article <472@unirot.UUCP>, samperi@unirot.UUCP (Dominick Samperi) writes:
...
> > I've since learned that the 3B2's do not format floppies in the same way
> > that the PC does. Does anyone know how floppies are formatted on the 3B2?
...
> Recently, I tried to read some 3b2 floppies on Microport SV/AT V2.2.2L on
> my AT clone.  I was able to read the first track...
...
> with both the 360k floppy and the 1.2mb floppy I got read errors when the
> floppy stepped to the next track.
...

I think the 3b2 machines use quad-density (720 kbyte) floppies. 
This probably means that the track spacing is different, which
explains why it's hard to step beyond track 0.  If this is the case,
then there is also a different media (not dsdd, and also not hd)
required.

Anyone at AT&T care to comment?
-- 
Dave Levenson
Westmark, Inc.		A node for news.
Warren, NJ USA
{rutgers | clyde | mtune | ihnp4}!westmark!dave

crs@cpsc6b.cpsc6a.att.com (C. R. Seaman) (10/06/87)

In article <1737@killer.UUCP>, wnp@killer.UUCP (Wolf Paul) writes:
< In article <472@unirot.UUCP>, samperi@unirot.UUCP (Dominick Samperi) writes:
< > I just finished implementing a UNIX-compatible cpio for DOS on a PC, and
< > I hoped to use it to transfer files via floppy disks between my AT
< > and AT&T 3B2's [ ... ]
< > I've since learned that the 3B2's do not format floppies in the same way
< > that the PC does.
< 
< Two pieces of information on this:
< 
< 1. Rich Salz recently posted Emmet Gray's MTOOLS package to comp.sources.unix.
< This package will allow any UNIX system to read MS-DOS diskettes, provided
< that the UNIX system has a floppy drive physically capable of reading the
< MS-DOS formats (i.e. 40 Tracks @ 9 sectors for 360K disks). This would allow
< you to copy a DOS file to the 3B2 or vice versa, and then unpack it with the
< appropriate archiver.

The *other* requirement for this system of tools is that the machine
already has a driver for the floppy drive that can read and write MS-DOS
formatted disks.  I know of no such driver for the 3B2. (That doesn't
mean there *isn't* one, just that I've never seen it ;-)).

< 2. The MKS Toolkit contains a dd(1) command, which permits reading and
< writing floppy disks without reference to the file system on the disk,
< i.e. it just reads/writes sectors in the proper sequence. With this command,
< I can write an archive to a floppy from DOS like this:
< 
<   find ... | cpio ... | dd of=a:
<
< [ ... ]

It would be much easier to use MKS Toolkit's version (or the UNIX compatible
version you have written) of cpio, create an archive file, use one of the
many file transfer protocols to transfer the file, then use cpio on
the 3B2 to extract the archive.  Cpio archives don't need to be written
to a raw device.  They can just as easily be written to files, which
may (or may not) be more 'portable'.

< Wolf Paul
< ihnp4!killer! [ mcomp! ] wnp

-- 
Chris Seaman            |    o\  /o
crs@cpsc6a.att.com <or> |      ||         See "Attack of the Killer Smiley"!
..!ihnp4!cpsc6a!crs     |   \vvvvvv/     Coming Soon to a newsgroup near you!
                        |    \____to 

tr@wind.bellcore.com (tom reingold) (10/08/87)

In article <418@cpsc6b.cpsc6a.att.com> crs@cpsc6b.cpsc6a.att.com (C. R. Seaman) writes:
$ [...]
$ It would be much easier to use MKS Toolkit's version (or the UNIX compatible
$ version you have written) of cpio, create an archive file, use one of the
$ many file transfer protocols to transfer the file, then use cpio on
$ the 3B2 to extract the archive.  Cpio archives don't need to be written
$ to a raw device.  They can just as easily be written to files, which
$ may (or may not) be more 'portable'.
$ [...]
$ Chris Seaman

Yes, that's easier.  But it misses the point.  Using a serial port, as
I think you are suggesting when you say "transfer the file", is slow.
A floppy disk drive goes MANY times faster than 9600 bps or 19200 bps.
When there are many files or large files or both, a serial port is
very impractical.

Another problem is that the cpio file could be bigger than the space
you have on a disk.

We really need something here.  I think we will soon have it, because
the need is so real.  Mostly, we need a floppy drive that connects
easily to Unix machines and we need a device driver on which to run
the software posted by Rich Salz.

Tom Reingold 			INTERNET:       tr@bellcore.bellcore.com
Bell Communications Research	UUCP: 		<backbone>!bellcore!tr
435 South St room 2L350		SOUNDNET:	(201) 829-5119 [work]
Morristown, NJ 07960				(201) 287-2345 [home]