This is the mail archive of the
gcc@gcc.gnu.org
mailing list for the GCC project.
Re: Libjava failures status
Jan Hubicka writes:
> > On Sun, Mar 10, 2002 at 01:06:36PM +0100, Jan Hubicka wrote:
> > > I plan to add symetrical check verifying that we do not discover
> > > trapping instruciton in the middle of basic block...
> >
> > If you do this, you'll need another representation for
> > instructions that, examined individually, may trap, but
> > global information has determined that it cannot.
On a somewhat related subject:
I remember a while ago I wrote a patch to the Java front end that
marked MEMs as either potentially trapping or not, and as far as I
could tell this information was preserved throughout all optimization
stages. This was pretty easy to do, and eliminated about of half the
instructions from being treated as potentially trapping.
Andrew.