> Ok, I posted about this all the responses came back asking me if
> I ran lilo after installing the kernel. Well, I don't use lilo, so no I
> didn't run it. One person asked if I tried dd'ing the kernel to a floppy
> and booting... that worked... But, not the solution.
>
Telling us what you do use would help, loadlin ? bootlin ? erm, ethernet
remote boot ? give us a clue ....
> I started looking for what changed, and found that 4 lines were
> added to linux/arch/i386/boot/bootsect.S in the _main: routine
>
> nop
> jmp over_magic
> .byte 0xF0
> over_magic:
>
Hmm, as I understand it, this makes the start of the kernel image (which
bootsect.S goes at the start of) have a magic number for file of the
assembled <nop> <jmp> <addr of over_magic> <0xf0> (i think that's right,
90 EB <^A> F0 ?), but that shouldent be a problem.
> When I commented out these lines the system worked properly. Is
> this a problem with non-lilo booted machines?
>
Looks like it, but what do you boot with, it could be just that ....
> Julie
>
Bryn
--
PGP key available from any good server \ Overload -- core meltdown sequence
1024/EB275221 1995/09/05 Bryn P. A. Jones | initiated.
<bpaj@gytha.demon.co.uk> (use this key) / This space is intentionally left
fingerprint: 83 AF D8 65 BC 52 74 31 | blank, apart from this text ;-)
67 1A 5F 31 F5 A0 9C 82 \____________________________________