Re: objdump -k -q -o 0xC0100000 in 2.1.7 (fwd)

babydr (babydr@nwrain.net)
Sat, 9 Nov 1996 15:22:17 -0800 (PST)


Forwarding the below for further Real GURU examination. JimL

---------- Forwarded message ----------
Date: Wed, 6 Nov 1996 07:19:54 +0100 (MET)
From:
To: babydr <babydr@nwrain.net>
Subject: Re: objdump -k -q -o 0xC0100000 in 2.1.7

On Tue, 5 Nov 1996, babydr wrote:

> >
> > Why does this work for everyone in the world but me?? Me=dipstick??
> > I can hand-make a kernel with other than 0 size if I change the second
> > objdump to match the start address of the link.
> >
> > Thanks in advance,
> > Mike
>
> By chance are you attempting to compile NCR53c8xx & AHA154x
> in the same kernel ? Even have AHA154x as a module ?
>
> A while back I sent off a note about trying to do this &
> get an output (using 'make bzdisk') of
> 0+1 data in
> 0+1 data out
> for the 'dd' of the boot image.
>
> This sound simular, JimL
>
No, I'm only using the aic7xxx driver for an aha2940.

I'm neeeeeeeeearly sure that the Makefile in arch/i386/boot is wrong on line 28
I think it should be
$(OBJDUMP) $(OBJDUMP_FLAGS) -o $(BZIMAGE_OFFSET) and not
$(OBJDUMP) $(OBJDUMP_FLAGS) -o $(IMAGE_OFFSET)
I do know that changing it to BZIMAGE_OFFSET produces a bootable kernel - I'm
using it right now (2.1.7). I'm puzzled by why this works for everyone else.
<snip>
mikeg:# objdump --version
GNU objdump version 2.7
mikeg:# ld --version
ld version 2.7 (with BFD 2.7.0.3)
I think that these are the latest/greatest.

Cheers,
Mike