FC4t2 no good without LILO

[Jeff Spaleta]

On 4/14/05, Mike Bird <mgb-fedora at yosemite.net> wrote:
> Run fdisk -l.  You'll find a partition marked active.  LILO chained to
> the boot loader in that partition.

Here's what i think... i think its not sufficient to look at the just
the first line of the od -c output. And i think you know that, and
your just wasting Paul's time.

I think there is a trivial header space thats not important to
bootloader operation.. and grub leaves that short header segment alone
when installing its stage1. Hell looking at grub's actual stage1 and
all i get are zeros from od -c /boot/grub/stage1 | head -1
0000000 353   H 220  \0  \0  \0  \0  \0  \0  \0  \0  \0  \0  \0  \0  \0

Here's a better approach
doing a diff of the first 446 bytes returned from od -c   of the mbr
boot.b and grub's stage1 file.  I'll break it down into several
steps..so for anyone playing along can get a chance to visually
inspect the output in actual files.

od -N 446 -c /dev/hda > /tmp/mbr
od -N 446 -c /boot/boot.b > /tmp/lilo
od -N 446 -c /boot/grub/stage1 > /tmp/grub

diff /tmp/mbr /tmp/grub
The only thing returned in this diff.. is the very short space at the
front of that 446 bytes.

diff /tmp/mbr /tmp/lilo  
returns essential the whole recorded output. Indicating to me that
even though the first few bytes in the mbr space were not overwritten
by grub.. the bulk of it was.. and that the first few bytes are not
important to grub operation.

If Paul is still paying attention to this thread, he might want to
compare his results to mine. Just to put the last nail in the coffin
over the somewhat silly idea that lilo is actually magically
chainloading.

-jef"is getting tired of using his powers for good"spaleta