> That said, I do agree that your inability to write a disklabel is
> wrong.  A fix I'd rather see, though, is for disklabel to stop
> worrying about whether the RAW_PART is a nice # of cylinders. Its
> size should be compared with the drives true size.

To clear up this misconception, disklabel is not checking whether the
raw partition is "a nice number of cylinders"; the only disklabel code
that's fussy about cylinder boundaries is inside an #ifdef notdef.

The problem is that disklabel's current text representation of a label
does not include a d_secperunit field, so disklabel computes
d_secperunit ("the drives true size") by multiplying d_secpercyl and
d_ncylinders.  This is clear not always accurate, and I've already
sent mycroft a patch to include d_secperunit in the text
representation of the disklabel.