Cleanups for the m68k backend

[Bernardo Innocenti]

On Tuesday 08 July 2003 09:47, Gunther Nikl wrote:

 > On Tue, Jul 08, 2003 at 07:36:03AM +0200, Bernardo Innocenti wrote:
 > >  Every obsolete feature you drop opens the opportunity for more
 > > cleanups.
 >
 >   True, but I won't call MIT syntax obsolete.

 It's not obsolete. It's just useless when no supported target
depends on it. Since it's useless, we have the opportunity to
declare it obsolete and drop it from GCC.

 > >  It's too bad, but I'm starting to think that no single person knows
 > > enough about all the environments where the m68k back-end is being
 > > used to be able to clean the whole back-end properly.
 >
 >   Quite likely given the fact that most m68k ports haven't been
 > touched since a long time. Since a lot of dead configurations are
 > already gone the m68k backend should be much more manageable now.

Not as managable as most developers would like it to be. There are
many areas where the m68k back-end shows its age and makes things
difficult to work with. The most invasive thing is that MOTOROLA/MIT
dual syntax.

I'd like to do something about this, so that the next developers looking
at that code in the future won't think GCC is crap.

Please note that I won't insist in dropping MIT syntax if it happens
to be badly needed by some mainstream targets. Within this thread we've
found out that some of the remaining targets could switch to the Motorola
syntax in no time. There's still one open issue about OpenBSD. Let's see
if we're lucky.

 > > I was quite surprised from the include layout when I first looked at
 > > it. I had a very hard time before I figured out that m68020-elf.h
 > > was being included last in my m68k-elf config and it was
 > > redefinining symbols already defined eariler by other m68k headers.
 >
 >   Then I would suggest to start cleaning up this first and tackling
 > other problems later.

 The problem here is that I don't know how it's supposed to work for
all possible targets. What does m68020-elf.h mean? Who introduced it
and for what reasons? Digging into this problem is something that's
best done by long-time GCC developers like you.

-- 
  // Bernardo Innocenti - Develer S.r.l., R&D dept.
\X/  http://www.develer.com/

Please don't send Word attachments - http://www.gnu.org/philosophy/no-word-attachments.html