🚀 go-pugleaf

In article <240@intelca.UUCP> danl@intelca.UUCP (Dan Lau) writes:
>I just bought a 10M hard disk for my IBM/PC, but after I formatted
>it, format reported a size of 10584064 bytes total available instead
>of the usual 10592256 that is reported on a PC/XT.  What is wrong?
>The particulars are:
>
>	IBM/PC with ROM BIOS dated when original PC/XT introduced
>	IBM/PC-DOS V3.1
>	OMTI Hard disk controller 5510-3
>	CMI 5412 drive (asme drive as used by IBM, 306 cylinders, 4 heads)
>
>FDISK was used to install one DOS partition using the whole disk.  No bad
>sectors were reported when doing the hard and soft format.
>
>Does anyone know why I am missing 16-sectors (2 clusters of 8K each).

"Elementary, my dear Watson..."  I'll bet you were using DOS2.1 on the
XT and just now switched to DOS3.1 when you FDISK'ed and formatted the
new drive.  Every release of DOS fiddles some parameter or another and
3.1 is no exception.  The culprit here is FDISK.

FDISK's task is to create and maintain "partitions" on the fixed disk.
It does this by storing a special boot program and partition table on
the first sector of the disk (ie - cylinder 0, head 0, sector 1).  In
the table are structures which describe the starting cylinder and
ending cylinder numbers for up to four separate areas, or OS'es, only
one of which may be bootable at a time.  Although the partitions may
be specified as beginning and ending at an arbitrary sector, FDISK
only allows the user to specify partitions to a cylinder boundary
(remember: cylinders are the "detents" that the mechanism carrying the
heads may step to, "tracks" is the product of "number-of-heads" times
"number-of-cylinders"; eg: the XT standard drive has 306 cylinders and
four heads, giving it a total of 1224 tracks).

In DOS2.1, if you specified to FDISK "gimme the whole disk", it would
set up the partition slot number 4 with the beginning of the disk as
cylinder zero, head zero, sector 2 and the end of the disk as the last
sector on the last track minus one (the last track is reserved for
diagnostics to scribble on).  Now the value for the start of the
partition disobeys already the rule we said about starting on an even
track boundary (eg: given cylinder, given head, sector 1) because the
partition and boot block lives on the first sector.  I suspect some
purist somewhere (and perhaps some confused software) convinced the
writers of DOS3.[01] to change that.  So they did.

Now, partitions begin at a minimum of cylinder zero, head ONE, sector
one, which means you lose 16 sectors (the remainder of track zero after
the boot block).          **********

A complete description of the contents of the "Boot Record/Partition
Table" is found in Appendix G of the PC/XT Technical Reference Manual.


Bruce Walker     {allegra,ihnp4,linus,decvax}!utzoo!aesat!bmw

"I'd feel a lot worse if I wasn't so heavily sedated." -- Spinal Tap