This is the mail archive of the
gcc-patches@gcc.gnu.org
mailing list for the GCC project.
Re: [patches] Re: Recognize cpu varaints take III
> >
> > > On Wed, Dec 12, 2001 at 09:37:18AM +0100, Jan Hubicka wrote:
> > > > ! { "mmx", MASK_MMX | MASK_MMX_SET, \
> > > > ! N_("Support MMX builtins") }, \
> > > > ! { "no-mmx", MASK_MMX_SET, \
> > > > N_("Do not support MMX builtins") }, \
> > >
> > > Ok, except that you must also have a no-mmx, -MASK_MMX entry so that
> >
> > Not ok! The win32-based (i.e. cygwin) builds are all broken now,
> > because MASK_DEBUG_ADDR now conflicts with win32's MASK_WIN32. It
> > also conflicts with DJGPP's MASK_BNU210 and dgux's MASK_STANDARD.
> Oops, I totally forgot about the target specific flags used there.
> Do we still have room for that? I will try to fix that shortly.
Oops, no room for in MASK - we do have more than 32bits used.
Was this ever hit in some other backend? What is the most convenient
sollution to this situation?
Honza
>
> Honza
> >
> > MASK_DEBUG_ARG conflicts with cygwin's MASK_CYGWIN and dgux's
> > MASK_NOLEGEND.
> >
> > MASK_INTEL_SYNTAX conflicts with win32's MASK_WINDOWS and dgux's
> > MASK_EXTERNAL_LEGEND.
> >
> > MASK_NO_RED_ZONE conflicts with win32's MASK_DLL and dgux's
> > MASK_IDENTIFY_REVISION.
> >
> > MASK_64BIT conflicts with dgux's MASK_WARN_PASS_STRUCT.