On Mon 19 Oct, Ignatios Souvatzis wrote:
> On Mon, Oct 19, 1998 at 01:55:36PM +0100, Peter Teichmann wrote:
> > How do I have to mount an ADFS partition wich is on IDE drive 0?
> > I guess it is something like "mount -t filecore /dev/wd0? /mnt".
> > But which wd0?. wd0a, wd0b, wd0e are the NetBSD partitions on that
> > drive. But which name has the ADFS partition? And why are the NetBSD
> > partitions not numbered wd0a, wd0b, wd0c? 
> 
> the "c" partition is used for the "all of the NetBSD part of the disk."

NO. Under the arm32 port, the c partition is used for ALL of the disk,
WHATEVER is on it.

> On machines where this is different (currently, all using MBR partitions),
> the "d" partition is used for "all the disk".

AH. This message is on port-arm32. The arm32 port uses partition c to
refer to the whole disk. There is no special partition that refers
only to the NetBSD part of it.

The d partition has no special meaning, but on Risc PCs commonly is used
to point to a filecore partition, which is always at the start of the disk.

On Tue 20 Oct, Dr. Stephen Borrill wrote:
> Surely the ADFS map on the wd0c refers to only the ADFS portion and
> hence partition wd0c can be used as the non-ADFS bit will be
> ignored? Certainly, I've never had any problems using mount_filecore
> /dev/wd0c /mnt.

As ADFS puts the filecore partition at the beginning of the disk,
mount_filecore successfully finds an fs on /dev/wd0c. However if I
(or anyone else) gets around to writing a read/write version, doing
this would be exteremely dangerous, as the filecore fs could corrupt
your whole disk, not just the filecore partition.

Anyone using a disk with multiple filecore partitions (ie with some third
party IDEFS's for RISC OS) can access these by creating a suitable
disklabel which has partitions pointing to the other filecore partitions.
The second one is usually immediately after the first, and the third
immediately after the second etc. However you will quickly run out of
partition labels (at h).

	Andrew