5 May
2005
5 May
'05
6:33 a.m.
On 5/4/05, Dennis Boone <drb at msu.edu> wrote:
Oh really? Odd. I recall doing that for SCSI but not for ESDI - I'm
pretty sure I needed to tell the BIOS there was /some/ drive there,
and at least a 20MB to be reasonably sure of getting all of track 0
and the partition table and so on without splitting that track over
several nonexistent cylinders. Mind you, whether my dim recollection
of type 20 being true is anybody's guess, it's so long ago now. This
entire bit could be false memory syndrome. :?)
--
Liam Proven
Home: http://welcome.to/liamsweb * Blog: http://lproven.livejournal.com
AOL, Yahoo UK: liamproven * ICQ: 73187508 * MSN: lproven at hotmail.com
I think I handled a 600 in a Novell server once.
Oooh! [Envy] ;?)
It's an
Int11 device, just like an ST-506 drive. The BIOS needs
to be configured with the right number of heads, cylinders and
sectors-per-track, tho' picking type 20 or so will usually give you
enough to boot and read the true settings off the drive itself. They
will probably need to be low-level formatted if moving them from
one controller to another. DEBUG and then G=c800:5 is what I dimly
recall for this, tho' in later years I used CheckIt or even SpinRite
to do this.
I'm inclined to disagree, though I'm hardly the expert.
Specifically, I recall having trouble setting up several ESDI drives
in servers. The fix was to tell the machine BIOS the disk didn't
exist, then let the ESDI controller BIOS work out geometry.