[Gmecm] motorola 68k / cpu32 assembly

[Ant]

Hi Jim,

That sounds likely. I'll post back when I know for sure.

Thanks,
Ant

On 21/03/2012 7:16 PM, Jim Sturcbecher wrote:
> Hi Ant,
>
> I hand wrote code for the 68k many years ago and had a similar problem.
> Because the 68k is a 32bit processor, I had to ensure that all code and
> data sections in my source started on a "word" boundary. I think your
> linker is trying to do this for you and if your compiled code is not
> aligned to a word boundary, then you will get this shift.
>
> Just a possibility, try this and see is it fixes your problem.
>
> Jim.
>
>
> -----Original Message-----
> From: gmecm-bounces at diy-efi.org [mailto:gmecm-bounces at diy-efi.org] On Behalf
> Of Ant
> Sent: 21 March 2012 09:26
> To: GM-ECM
> Subject: [Gmecm] motorola 68k / cpu32 assembly
>
> Has anyone here attempted to assemble 32 bit motorola 68k code for an
> '0411 pcm?
>
> Im trying to use the sourcery codebench build of the gnu tools, but I
> think the linker is buggy.
>
> If I use m68k-linux-gnu-objdump -r to view the relocations in the .o
> file they check out correctly. But when I use the gnu linker to resolve
> the relocs and spit out a .bin file, they are all off by 2.
>
> Its not a linker configuration error as if I offset the load address
> of .text too offset addresses by 2 lots of other things break.
>
> tool versions are:
>
> $ m68k-linux-gnu-as --version
> GNU assembler (Sourcery CodeBench Lite 2011.09-22) 2.21.53.20110905
>
> $ m68k-linux-gnu-ld --version
> GNU ld (Sourcery CodeBench Lite 2011.09-22) 2.21.53.20110905
>
> I'd be interested to hear from anyone who has written code by hand and
> assembled with these tools.
>
> Thanks,
> Ant
>
>
>
> _______________________________________________
> Gmecm mailing list
> Gmecm at diy-efi.org
> http://lists.diy-efi.org/mailman/listinfo/gmecm
>