Re: function arguments offset computation with -fomit-frame-pointer

[Hans-Peter Nilsson <hp at bitrange dot com>]

On Wed, 9 Jul 2003, Bernardo Innocenti wrote:
>  > >  Could I use a simulator of some kind?
>  >
>  > Yes, this is IMHO the easiest way to do it (once code is in
>  > place that is; after setting it up the first time).
>
>  I've seen it's rather complicated. Too bad gcc/binutils/gdb don't
> share a single CVS tree.

It's quite simple if you just follow the simtest-howto.html
instructions.  They're easily scripted; just copy from that
page.  That is, for existing simulator setups.  For nonexisting
ones, it wouldn't help very much with a single CVS tree anyway.
:-)  (There are "only" two trees -- not that a single tree isn't
the goal; the reasons it officially isn't are political.)

>  > There's a newlib and a libgloss port for m68k, and it seems the
>  > necessary simulator low-level bits have been written, but I
>  > don't know for what simulator, and I see no m68k simulator in
>  > sim/ or sid/.
>
>  I see. There's an interesting ColdFire emulator which is even able
> to run uClinux:
>
>    http://www.lightbox.org/coldfire/

Looks promising.  Care to do it, or to outline such a project on
projects/beginner.html?

> But I guess making it work with gcc's test infrastructure requires a
> lot of integration.

For reference: You mean making it work with dejagnu, newlib and
libgloss.  GCC doesn't have much to do with this part.  And
having done such work before, I kindly disagree it'd be much
work, but then again I'm not the one who'll do it for this
setup. :-)

> It would really be too far away from my original
> project ;-)

Maybe, but very useful and valuable, particularly as a way to
make sure the coldfire support remains(?) working. 1/2 :-)

brgds, H-P