>> 1. getting disklabel handling code out of disk device drivers
> Could you please show me where in the code the disk drivers have
> knowledge of the disklabel?  They have knowledge of a structure in
> memory which tells them [...] that partition X starts at point Y and
> has length Z.  They also know a max # of partitions, and that
> partition RAWPART is special - it bypasses the start-length table.

It seems to me you just answered your own question.

Why should that code you just described be either (a) duplicated in
every disk driver or (b) in the same place as disk-specific code to
talk to the hardware?  I can't see any reason for either one, and good
reasons to avoid each.

My current preferred solution is the labelfs/partition-driver solution
I mentioned in a message I sent yesterday.

> The drivers are always going to need to know some information about
> the partition layout,

*Some* piece of the kernel will...but there's no reason it has to have
either of the properties (a) or (b) above.

					der Mouse

			       mouse@rodents.montreal.qc.ca
		     7D C8 61 52 5D E7 2D 39  4E F1 31 3E E8 B3 27 4B