[ECOS] ELF to AXS

Andy Simpkins Andy.Simpkins@TardisMobile.com
Tue Mar 27 00:54:00 GMT 2001


Like Carlos, I too am working with the ARM tools (SDK and ADS with
MultiICE).  This is because it is the platform used by the majority of the
application software that we already have.  What we as a company did not
have was an RTOS.  We chose eCos.

There is no compatibility with the elf format generated by eCos and that
understood by the ARM tool chain.  This caught us out as well!

I have succeeded in circumnavigating my problem (of having object code only
under ADS) by simply calling the routines at there location in memory... not
nice but it works...

Andy

Below is a thread from the newsgroup if it is of interest

Andy Simpkins <andy.simpkins@tardismobile.com> wrote in message
news:<982578547.24806.0.nnrp-08.d4e428b3@news.demon.co.uk>...
> Ok, the next logical question has to be has anyone recompiled eCos under
> ADS?  And if so how did you cater for the eCos configuration tool - I
realy
> could do without interpreting all that CDL by hand!
>
> Andy
>
>
> Wilco Dijkstra <Wilco.Dijkstra@NOSPAMarm.com> wrote in message
> news:3A8D5C39.C548F557@NOSPAMarm.com ...
> > Andy Simpkins wrote:
> >
> > > Can anyone suggest a possible solution please?  (other than attempting
> to
> > > rebuild the eCos library under ADS).
> >
> > I can't comment on the particulars of the problem - you could try
> extracting
> > the objects from the library, and then I believe it should link at least
> with
> > ADS 1.1.
> >
> > However, be aware that today it is impossible (except in trivial cases)
to
> > mix objects from 2 different compilers. For example, the calling
standard
> that
> > the GNU compiler uses is different from the ARM compiler in many
> (non-trivial)
> > ways. If you're an expert in both compilers you can get it to work for
all
> but
> > the corner cases. This will change once most compilers adhere to the ARM
> EABI.
> >
> > So at the moment it's best to recompile eCos under ADS...
> >
> > Wilco
>
>




More information about the Ecos-discuss mailing list