This is the mail archive of the
gcc@gcc.gnu.org
mailing list for the GCC project.
Re: Recent changes to cse.c
- To: law at redhat dot com
- Subject: Re: Recent changes to cse.c
- From: Denis Chertykov <denisc at overta dot ru>
- Date: 24 Jul 2001 00:11:40 +0400
- Cc: Richard dot Earnshaw at arm dot com, Richard Henderson <rth at redhat dot com>, Mike Lerwill <mike at ml-solutions dot co dot uk>, gcc at gcc dot gnu dot org
- References: <18855.995914109@localhost.localdomain>
law@redhat.com writes:
> In message <200107180903.KAA09819@cam-mail2.cambridge.arm.com>you write:
> > > IMHO, cc0 should be exposed during final because that's the only time
> > > we actually care about the contents of cc0 (for redundant compare/test
> > > elimination).
> >
> > We have mechanisms now which make HAVE_cc0 unnecessary (compare-and-jump
> > patterns). Wouldn't it be easier to just get rid of HAVE_cc0 entirely
> > from the compiler and convert those few remaining ports that use it to the
> > new code?
> I'm not familiar with this support. But it sounds like the model I've got
> for compare/jumps in general.
What must I read to do a right conversion from HAVE_cc0.
Denis.