[ECOS] Re: Réf. : Re: [ECOS] Difference between generated binary files from Linux and Windows (cygwin)

Nick Garnett nickg@ecoscentric.com
Wed Jul 28 21:08:00 GMT 2004


e.coullien@faiveley.com writes:

> I launched the command "powerpc-eabi-objdump -D" on the two .elf files. See the
> attachment file.
> The two codes seem to be the same but there are some offset or adress which
> differs.

It is not a good idea to post 1MB+ attachements to this list. With
luck the list server will have bounced it. But Gary and I still got it
and you should not be sending such large attachements unsolicited. If
either of us were on a dialup connection it would have been a nasty
surprise. Various spam filters might also be upset by it.

My intention was to try and get you to investigate and find the
solution yourself, rather than send it all to us.

However, since it's here, I took a look. The only real difference
between these executables seems to be the order in which the data
sections are layed out, which results in differing offsets in the
code. This is usually a result of the files in libtarget.a being in a
different order, which can happen if some part of eCos is rebuilt. If
this is happening with builds from scratch then I have no idea why the
build order would differ. However, it should not make any difference
-- it's arbitrary to start with and an arbitrary reordering will be
benign.

> 
> I try some other tests :
> If I build the kernel eCos on Cygwin and I compile the application on Cygwin, I
> obtain a different code than on Linux.
> but if I build the kernel eCos on Linux and I compile the application on Cygwin
> with this kernel, I obtain an identical code than on Linux.

This is entirely consistent with what I said above.



-- 
Nick Garnett                    eCos Kernel Architect
http://www.ecoscentric.com/     The eCos and RedBoot experts


-- 
Before posting, please read the FAQ: http://ecos.sourceware.org/fom/ecos
and search the list archive: http://ecos.sourceware.org/ml/ecos-discuss



More information about the Ecos-discuss mailing list