Re: [patch] Fix for libgloss/m68k

[Jeff Johnston <jjohnstn at redhat dot com>]

Kazu Hirata wrote:
> Hi,
>
> > Regarding problem 1. I was kind of hoping some m68k experts would pipe
> > up. Is there a workaround for the halt instruction? Is there a
> > compiler architecture flag that could be checked to see if the halt
> > insn is available? (similar to the situation we recently ran into with
> > the ARM bx instruction).
>
> A patch I posted a few days ago
>
>   http://sources.redhat.com/ml/newlib/2006/msg00965.html
>
> fixes the problem.  Specifically, I have libgloss/m68k/configure check
> for __mcoldfire__.  If that is defined, then I build the ColdFire
> portion of libgloss/m68k.  In terms of actual script, I do something
> like:
>
> AC_MSG_CHECKING([target cpu family])
> DO=m68k
> AC_PREPROC_IFELSE([#ifndef __mcoldfire__
> 		#error we are not coldfire
> 		#endif],
> 	       DO="cf",)
> AC_MSG_RESULT($DO)
> AC_SUBST(DO)
>
> And then I use DO in Makefile.in to build an appropriate portion of
> libgloss/m68k.

FYI.  I just checked in the patch so you can try it out.

> This approach should fix the problem you are having.  Thoughts?
>
> Thanks,
>
> Kazu Hirata