i386/70525: [boot] boot0cfg: -o packet not effective

Brian Candler B.Candler at pobox.com
Tue Dec 5 00:27:15 PST 2006


On Tue, Dec 05, 2006 at 08:19:44AM +0100, Remko Lodder wrote:
> > Are you saying that the FreeBSD MBR can only boot operating systems which
> > were installed at the time that the MBR was installed?
> 
> No I am not stating that, I was just theorising there, I am not aware on
> the internals of the MBR, but I could imagine that it can only boot systems
> that have been installed (since then there is information about the operating
> system and the boot 'sector' on the partition, prior to that
> it is unfilled and the system cannot tell whether it is going
> to be used to boot or not.
> 
> > 
> > I don't think this is the case, because the MBR correctly detected both
> > partitions, and offered them to me:
> 
> Well, are you sure that the F2 here implies the same partition?

Yes, because:
(1) before installing OpenBSD, it only offered F1
(2) after installing OpenBSD, it offered F1 and F2 (but F2 just went beep)
(3) after reinstalling the MBR, it offered F1 and F2 (and F2 worked)

> > All it has to do is to load the boot sector from that partition, and jump to
> > it. I'm presuming that this requires the new 'packet' BIOS call when this
> > sector is located above the 1024-cylinder mark.
> 
> Imo all you need to do is add the new partition to the mbr bootblock and it will
> start.

Add it where exactly? There is the partition table. There is the boot code.
There is a byte containing options flags. As far as I know the MBR does not
contain a duplicate boot table.

I guess this discussion is all my fault for not having taken a copy of the
MBR before and after boot0cfg, as it would have been easy to prove that the
only difference was the flags byte.

Regards,

Brian.


More information about the freebsd-i386 mailing list