PPC float context only 4 Bytes per register?

[Sergei Organov]

Dear Eric,

I didn't blame anybody for anything. I also don't complain. I just attempted
to answer Joel's question. If I need to use new exception processing, I'd fix
it in few minutes, but old exception processing still works fine for me, so
why bother?

BR,
Sergei Organov.

"VALETTE Eric" <valette at crf.canon.fr> writes:
> Sergei Organov wrote:
>
> > Joel,
> > Code in old exception processing has two variants selectable at compile
> > time:
>
> > 32 bits and 64 bits registers. I don't know any PPC variant that has 32
> > bits FPU registers though. Moreover, it seems that with old exception
> > processing none of PPC variants supported by RTEMS actually used 32 bits
> > registers. I can't imagine why in new exception processing code only
> > unused 32 bits code was left.
>
>
> Only those who make nothing makes no mistake :-) So It is probably my mistake
> when developping new_exception_processing code...
>
>
>
> But I guess it
>
> takes about 15mins to copy old_exception_processing/cpu_asm.S code code to
> new_exception_processing/cpu_asm.S.
>
>
> I have no time to do that right now. So either you do it or you stop
> complaining given the amount of time it represent...
>
>
> I do not use FPU in my application so never detected the bug... I wrote
> the new exception processing code that offers much more feature and a cleaner
> API, you can spend some time correcting the few bits left...
>
>
>
> --
>     __
>    /  `                          Eric Valette - Canon CRF
>   /--   __  o _.                 Product Dev. Group Software Team Leader
> (___, / (_(_(__                 Rue de la touche lambert
>                                  35517 Cesson-Sevigne  Cedex
>                                  FRANCE
> Tel: +33 (0)2 99 87 68 91       Fax: +33 (0)2 99 84 11 30
> E-mail: valette at crf.canon.fr    http://www.crf.canon.fr