[Thus spake Matthias Drochner ("MD: ") 9:13pm...]

MD: bouyer@antioche.eu.org said:
MD: > The disklabel is at the beginning of c:, not d:. This may be enouth to
MD: > make sure that c: is the NetBSD partition of the disk.
MD:
MD: The disklabel is found by using the offset found in the MBR - at
MD: that point it shouldn't matter whether a c: entry points to
MD: the same place or not.
MD: Again - please excuse if I'm missing something, it was a while
MD: ago when I looked into this.

I think I have to agree:  If we enforce complete unchangeability of
the 'c' partition information, including the FS type, this means we
cannot use 'c' as a partition at all.  I've never seen an implementation
which so denies the use of the 'c' partition; even Solaris allows it
if you tweak the flags properly.

That said, c: _should_ encompass the entire NetBSD portion of the disk.
Whether 'a' points to the same place as the start of 'c' is totally
irrelevant.

[if I'm off target here, ignore me]

MD: best regards
MD: Matthias


				--*greywolf;
--
Woof...