On Wed, 13 Oct 2004, thilo wrote:

> 16 partitions:
> #        size    offset     fstype  [fsize bsize cpg/sgs]
>  a:  10239453  16465491     4.2BSD   1024  8192    93   # (Cyl. 16334*-
> 26492)
>  b:    524288     16065       swap                      # (Cyl.   15*- 536*)
>  c:  15925138    540353     4.2BSD   4096 32768  1184   # (Cyl.  536*-
> 16334*)
>  d:  26704944         0     unused      0     0         # (Cyl.    0 -
> 26492)
>  f:     16002        63 Linux Ext2      0     0         # (Cyl.    0*- 15*)


> and after booting into 2.0
> wd0c  shows and mounts as wd0a  (wd0a is still the same as before)

Here's me thinking aloud:

  I notice that 'c' is placed before 'a', numbers-wise.  It sort of makes
sense that the range that is 'c' would get picked as the NetBSD part if
there are two DOS partitions with the same (NetBSD) sysid and it is
specified first (in the DOS partition table).

  And then 2.0, due to some changed logic or reordering of the code, picks
the other DOS partition - which coincidentally is also 'a' above - to be
the NetBSD part; and then 'c' obviously becomes a duplicate of 'a'.

  Do I understand the partition handling right?  And if so, what caused
the change in interpretation?


Magnus