23 Nov
2006
23 Nov
'06
9:29 a.m.
Chris M wrote:
when I originally setup this drive on an old pentium,
which couldn't support it's capacity w/o an overlay
either, it formatted it to the capacity it was able.
Shouldn't dos on the 486 do the same?
remember that pc controllers for hard drives, and later IDE drives
get their capacity from what you tell them.
when a 486 bios boots up, it will use the HD access definition that
the bios supported. It may be compatable with the original AT only
and only go to ~500mb, or may have bios that will work with the
# of sectors / trk and # of cylinders maxed out and go to ~2.2gb
total (some called it LBA, some bios other things).
The trick is to get a boot written that will not remap the boot region
when you change methods of access.
you may well not have a bios with can write 63 sectors / track
and init the disk to that. If so, you will need a boot that will
work with the lowest common denomentator, which is to map
to the 0:0.0 spot on the disk, then read a boot from the remainder
of what would be the first track, then it will have to be able to
remap and replace the :80 / :81 bios hard disk requests and
read the disk till the OS comes up. I thought that LILO installed
on the disk as the boot in early versions of linux handled this
problem. It may not be possible to have more than one
OS on the disk if you use Linux and LILO on a system,
since you cannot mix up the OS's if they don't share the
mapping of the cylinders, etc.
The 8gb mapping, 32gb mapping, and huge disk mapping
all use a different way of passing the disk address seek info
after the disk is booted by the bios with the smaller older
seek commands, but the whole mess has to be carefully
managed to be sure all the mappings are able to miss each
other during the boot process.
Jim