Its a hack, albeit a useful one. The offset to the BSD area of the
	disk (and disklabel) is in another data structure (eg: fdisk)
	and effectively available as a 'readonly' partition in the
	disklabel.

	This gives us the bizarre 'c vs d' schizophrenia within tools
	trying to access the raw partition on different ports.

	What does 'c' end up being used for on i386 that could not be
	equally well done with a 'whole disk' 'c'?

		David/absolute

                 -=-  "You can have my absence of faith"  -=-

On Mon, 16 Aug 1999, Dan Winship wrote:

> >         I've favour 'D' on the grounds its the next available partition,
> >         and the i386 'd' braindamage should go away eventually :)
> 
> So what is the right thing for a platform where disks may contain both
> NetBSD and non-NetBSD partitions? The i386 'c' vs 'd' concept is
> useful there.
> 
> -- Dan
>