[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: Partitioning

Bill Leach <bleach@bellsouth.net> wrote:

 > On Wed, 10 Dec 1997, Gertjan Klein wrote:

 > Obviously
 > these are judgement calls and opinions but when the original hard disk code
 > was written decisions were made concerning such things as sizes for device
 > storage parameters.  While what you have said about the cost of 10Meg HDs
 > and the like is true, that fact did not seem to influence others in such a
 > limiting way about how to deal with the matter.  More importantly, I
 > think, is that it has taken many years to finally to address this issue.

  The original harddisk code was written for (relatively) cheap
hardware.  SCSI harddisks, using block addressing, were of course
around, but much more expensive.  Nevertheless, even though the
partition table entries specify the location of partitions in cylinder,
head, and sector parameters, they _also_ specify them in logical block
numbers.  Using these numbers a harddisk of 2 TB (2048 GB) can be
described.  Note, again, that the partition table layout has nothing to
do with the BIOS.  The BIOS provided an interface for cheap hardware; if
demand would have been higher for better quality hardware, like built-in
support for SCSI drives, it would have been there.  Nothing in the PC
design prevents this - in fact, my BIOS directly supports (NCR) SCSI
controllers.  On top of that, the PC design allows for _really_ non
standard (for PC's, anyway) hardware to have it's own BIOS to take over
the standard BIOS calls.

 >> I'm not exactly sure what you think is the nightmare part of the
 >> original design (and frankly, I don't care).  There are a ...

 > And if you don't care then we are probably both wasting our time.

  No - because I am not trying to change your _opinion_ on PC hardware,
I am just trying to stop you from spreading misinformation about it.  I
realize very well that a lot of compromises have been made with PC
design over the years, there is enough to complain about - so if you
want to do that, go ahead, but get your facts straight.

 >> * There is a limited number of primary partitions available in the MBR.
 >> This limitation is no serious problem, as many modern OSes don't object
 >> to being installed in an extended partition (of which there can be as
 >> many as required).

 > Yes, many often incompatible workarounds exist.

  What do you mean with "incompatible workarounds"?  What's incompatible
about booting from an extended partition?

 > No there is nothing brain dead about partitioning a drive and I see no way
 > that anyone could conclude from anything that I have said that I think
 > otherwise.  It is the arbitrary decision to create the "tiered" partition
 > types (primary, extended, and logical) abstraction that I object to.

  Since the partition table resides in the MBR, with limited space,
_some_ limit had to be set to the number of entries in the table.  Four,
at the time, was a reasonable limit.  When the limit became - well,
limiting, MS introduced extended partitions - which is nothing other
than a way to arbitrarily expand the partition table.  It seems like a
reasonable solution to me.

 > These "modern BIOSes" have finally caught up with BIOSes of more than
 > twenty years ago.  Are you suggesting that had different decisions
 > concerning how to deal compatibly with the various limitation that were
 > arbitrarily built into the original design had been handled differently
 > that the PC would not be as popular or have such a favorable
 > performance/price ratio as it currently has?

  Yes.  (I don't want to get into this, though, because there is no way
to prove one way or the other).

 > I have "lost it".  In as much as I really do not wish to mislead anyone
 > then by "misinformation" are you talking about my assertions with respect
 > to the BIOS design (and indeed design evolution) upon the overall
 > filesystem design, or rather my (admitted) failure to even mention that
 > there are new BIOS designs that do not themselves impose this scheme, or
 > both?

  Your misinformation was that:

 - BIOS imposes the current partitioning scheme opon us, and limits the
number of primary partitions to four (not true - BIOS knows nothing
about partitions and doesn't care either).

 - DOS, Windows and OS/2 don't see other primary partitions than the one
they booted from (not true - DOS and Windows see other primary DOS
partitions just fine, and OS/2 won't even boot when they are present and
not "hidden").

 - fdisk /mbr will wipe out everything on the drive (wrong - it just
replaces (or installs) the MBR software without touching the partition

 - (A point I hadn't addressed yet:) loadlin uses BIOS calls for drive
access (wrong - it uses DOS calls, which can, but don't have to,
translate into BIOS calls).

  You _still_ don't seem to get that partitioning and BIOS have nothing
to do with each other, and that the BIOS is simply just a piece of
software - if the current interface is not longer sufficient, it is very
easy to change the BIOS to provide a new interface, or extend the
current one.  This has happened already - if it happened too late to
your liking, complain about the market forces that dictated the
development direction.  The older XT BIOSes were around 8 kB - current
BIOSes can be as large as 1 MB.  Things _have_ changed.

 > As I pointed out in the previous message, I have seen many PCs with
 > multiple primary partitions where only ONE primary partition is visible to
 > DOS or Windoz (or OS2 for that matter).

  This may be a silly question, but are these other partitions DOS
partitions?  If not, why do you expect DOS to natively support non-DOS
filesystems?  If they are, the only explanation I can give is that the
used boot manager (presumably OS/2's, but mine can do that too) hides
the other partitions.

[Hardware behaving sub-standard and incompatible]

  No argument from me here (other than that this, again, is not inherent
to PC design but to manufacturers selling hardware that is not up to
standard, presumably to make it cheaper).

 > In a sense all of this opinion matters not.  The PC is what it is and it
 > does seem to be evolving in the "right" direction even if it has taken
 > nearly 15 years to do make significant progress.

  You are right, all these opinions don't matter, and worse, are
off-topic.  I apologize for not sticking to the point I was trying to
make, which is not to refute your opinions about the PCs, but to correct
the errors described above.


Gertjan Klein <gklein@xs4all.nl>
The Boot Control home page: http://www.xs4all.nl/~gklein/bcpage.html

TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word "unsubscribe" to
debian-user-request@lists.debian.org . 
Trouble?  e-mail to templin@bucknell.edu .

Reply to: