This is the mail archive of the
gcc-patches@gcc.gnu.org
mailing list for the GCC project.
Re: spec20000 eon fix
> On Fri, Aug 03, 2001 at 10:28:48PM +0200, Jan Hubicka wrote:
> > What exactly goes wrong in Java? Can I apply similar trick?
>
> Perhaps. What will happen is that
>
> (set (reg tmp)
> (div (reg a) (reg b)))
>
> gets reloaded to
>
> (set (reg r1)
> (div (reg r2) (reg r3)))
>
> (set (mem (plus (reg sp) (const_int offset)))
> (reg r1))
>
> where the divide must still end the basic block, since it throws.
OK, I can take a look for the edges - what flag they get?
How can I recognize instructions that throws?
I will check find_basic_blocks to get the answers :)
Honza
>
>
> r~