Subject: Re: fdisk
To: port-i386 <port-i386@netbsd.org>
From: Andy Ball <andy.ball@earthlink.net>
List: port-i386
Date: 09/05/2006 05:50:24
Hello David,

   ASB> if I had a partition 20,859 sectors long starting at sector 17,
      > it ends at sector 20,875 but fdisk reports 20,876.

    DL> [1] The problem is that you want to say cylinder 1-12, not 1-11
      > /15/62 when the partition contains a whole number of cylinders.
      > And it is much easier to omit the numbers when they are zero!

I'm not convinced.  If I'm partitioning in cylinders (which I used to
do all the time, but PC/ATA geometry juggling makes that less
attractive) then I expect the end value to signify the end of the
cylinder mentioned.  In my example above, let's assume a geometry of
614 cylinders, 4 heads and 17 sectors
per track.  My partition runs from                Start    End   Size
cylinder 0 (excluding the boot track)             ~~~~~  ~~~~~  ~~~~~
to the end of cylinder 306. I would                   0*   306    307*
expect to see something like...

Start    End   Size          ...so that I could reasonably add another
~~~~~  ~~~~~  ~~~~~          partition starting from the cylinder
     0*   306    307*         after the one listed as the end of the
   307    613    307          first partition.  I think we already have
                              * as a convention for illustrating a
boundry somewhere within a cylinder (seen here because the first
partition starts after the boot track).  Why not use that to signify
other partition boundries that aren't cylinder-aligned?

- Andy Ball