Hello to Ale and Scott!

In message <Pine.SUN.3.91-941213.960916152851.2487A-100000@dust0.dur.ac.uk> Ale
wrote:

>   Anyway, you should get a
> non-adfs offset = 0001FB18
> RiscBSD partition @ 0001FB18
> message from disklabel.   
> 
> and the cumana software should say that your non RO area starts at 1290816

OK disklabel tells me:
non-adfs offset = 00157200
RiscBSD partition @ 00157200

The Cumana software tells me (after clicking on "Clear") that:
Start of non-RiscOS area: 1290240
Length of non-RiscOS area: 820571

The partitions are:
400 MB   Drive 4        cyl.    0-1422 according to bb_riscbsd
230 MB   Drive 5        cyl. 1423-2240 (output of bb_riscbsd, but corrected)
400.7 MB RiscBSD slice  cyl. 2241-3657

Hmm, 1290240 = &13B000 != &157200 is one thing to note...
If I use the value from the Cumana software, the result is 2240 cylinders
offset ie. cyl. 0-2239, this would mean I might even have one cylinder unused
between RiscOS and RiscBSD.
The value disklabel tells me can't be correct as this equals 2440 (or 2339 ?)
cylinders for RiscOS which would mean the RiscOS partitions are 702720 kB
= 686,25 MB large, which is not quite true.
I remember that I was told somewhere during the installation (of the alpha2 or
1.1-beta back in the old days) that the RiscBSD part started at cylinder 2440
because I made a remark about this in my notes.
As this is wrong I decided to ignore it and calculated the above value of 2240
as the starting cylinder for RiscBSD and used that. Versions of !SCSIMgr before
1.58 didn't tell the "Start of non-RiscOS area" anyway.

 In message <1894.199609161722@mortimer.cc.ic.ac.uk> Scott wrote:

> Looks OK. But the symptoms still suggest that the disklabel and the
> swap partition overlap. Could you send the disk geometry (as RiscOS
> sees it) and the cylinder offset in the filecore bootblock (bb_riscbsd
> should return it)

I'm not sure what you mean with "disk geometry as RiscOS sees it". You mean the
values the Cumana !SCSIMgr lists in the windows available when you select "Mode
Sense"? These are:

Under "Format Device Parameters"
Sector Per Track       120
Bytes Per Phys. Sector 512
Track Skew Factor       12
Cylinder Skew Factor    16
(Btw, what are the last two parameters?)

Under "Rigid Disk Geometry":
Cylinders 3658
Heads 6

RiscBSD says:
sd0: 1030MB, 3658 cyl, 6 head, 96 sectors, 512 bytes/sector

How come !SCSIMgr (1.58) claims it has 120 sectors per track?!? This also can't
be true as this would mean about 1286 MB - I definitely don't have that and
never had, even when the drive only had RiscOS partitions on it.
Anybody know the real values of a Conner CFP1080S? (But I still think the
RiscBSD values ie 96 sectors is correct, at least this almost fits the actual
size.)

Will we ever find the solution to this problem? ;-))

Markus