Eugenio Diaz wrote:
> I don't really understand what you are trying to achieve here.
That is obvious. It is my experience that a machine with no drives
on ide0 or ide1 (this is BEFORE the stanza IDE remapping) and using
the xosl boot manager will eventually collapse and become unuseable.
This process can be accellerated by installing and using grub instead
of lilo, then removing it in favour of lilo again. At this point
you will find that nothing will boot beyond LI even after running
/sbin/lilo again, in other words, the lilo secondary boot manager,
whereever it is, has been removed by grub and a new one can no longer
be written. There is no error message from lilo.
> I don't have any drive at the mobo controller (what would normally be ide0
> without the append string). The whole concept of ide0, ide1, ide2, and ide3 is
> a *linux* concept, not really pertaining to lilo or the bios. The append string
> is saved in the master boot record of the drive specified in the lilo.conf when
> you run lilo,
Huh? I presume you mean partition, not 'drive', and 'lilo.conf
stanza' not 'lilo.conf'.
> so there is no need to have a drive or a file system from where
> to read the append string; so why use the lock command?
So that the lilo deliberations are performed and the secondary boot
loader is placed on the remapped drives, otherwise a drive on IDE0
hda (before remapping) is required.
> Now, from what drive does the machine boots (or read lilo for that matter), is
> a whole other matter, but that's controlled by the bios, and has nothing to do
> with lilo or linux. I guess that an intelligent bios will search all drives on
> all ide channels until it finds one with a bootable master record, and then
> boots from there.
No. All the BIOS knows is to boot the first partition it encounters
in its search path which has its active flag set. It then reads the
MBR of that partition with no checks whatsoever and passes control to
it with no checks whatsoever.
--
Regards,
Ron. [au]
