[SunRescue] Quirks in SS10 floppy operation? Aw Pfutz..

[User Bobkeys BSD Bob the old greybeard BSD freak]

> Straight from the Sparc installation notes (install.sparc) for OpenBSD 2.8:
> 
> "Hardware the we do NOT currently support, but get many questions about:
> 	Multiprocessor machines (Will not boot correctly on multiple CPU sun4m)
> 	Audio driver for some sun4m machines (LX,ZX,ss10,ss20)
> 	Floppy driver for sun4m (can boot/install from ramdisk floppies though)"

Aw.. Pfutz....  I have been installing for so long on sun4c machines that
I did not read the details of sun4m bits.  Silly Me....(:+{{.....

It can boot fine on a single disk (openbsd), but not with two disks
(e.g. netbsd).  The drivers must not be picking up the disk changes
correctly.

> >3.  Boot/load of first floppy of NetBSD works fine, but change to
> >    second floppy to finish boot/load crashes and dumps the system.
> 
> I remember seeing some pretty stern warnings in the OpenBSD install notes
> about the installation being intolerant of floppies with defective sectors.
>  Maybe you've just got some duff disks?

True, but I have NEVER had problems with this on OpenBSD, with any of
my floppies, but the point is well taken.  I would expect it is the
driver issue that is the culprit.  Both NetBSD and OpenBSD are pretty
much the same driverwise, and I see it in both OS's.  FreeBSD handles
the things in stride on the old 486/33, though.  My rule of thumb is
that if the floppy won't format on FreeBSD, by 3 passes, it is frisbee
material.  Floppies that folks say are full of errors are usually not.
The tracks have to be rewritten sufficiently strongly (3 passes is
sufficient) to counteract any errors due to miswriting magnetically.
Hard errors are another problem.  Back in the early days when I did
a lot of 8 inch floppy stuff, I had to write special formatting utils
that would make multiple passes on a track to format.  That would
allow one to forcibly reuse otherwise dead floppies.  Remember how
expensive the old 8 inch DSDD things were.  I would go around and
collect up all the ``dead ones'' and laugh all the way to the bank.
It was easier on the peanuts computing budget back in those days.
Hard errors are really rather rare, but magnetic glitches due to
improper handling or improper use are rather common.  Slight
misalignments are also problematic.  But, I don't think that is
the case here.  It really does seem like a driver issue now that
you mention it.  You know, maybe it would be good to do a rewrite
of fdformat to handle multiple track write passes, to cure the
soft errors that occasionally crop up.

I will snoop around the Net/OpenBSD lists and see if anyone on the
porting crews knows of driver fixes.  OpenBSD 2.9 should be out
in a few days, as well as NetBSD-1.5.1.  My guess is that they
are not fixed, though, since I saw the problem on a beta snapshot,
too.

I still do a lot of floppy writing... guess I need to keep
the old SS1 or FBSD box handy, for a whiles longer...(:+}}...
 
Thanks

Bob