https://gcc.gnu.org/bugzilla/show_bug.cgi?id=91549
--- Comment #7 from Uroš Bizjak <ubizjak at gmail dot com> ---
(In reply to rdapp from comment #6)
> (In reply to Uroš Bizjak from comment #5)
> > This approach is not OK, you should use lp64 effective target instead of
> > -m64 option. Please see many examples in testsuite/gcc.dg.
> >
> > Also, x86 is a multilib target, described as { i?86-*-* x86_64-*-* }.
>
> Would this one do?
>
> diff --git a/gcc/testsuite/gcc.dg/wrapped-binop-simplify.c
> b/gcc/testsuite/gcc.dg/wrapped-binop-simplify.c
> index 44d85c04bfb..a5d953b46c7 100644
> --- a/gcc/testsuite/gcc.dg/wrapped-binop-simplify.c
> +++ b/gcc/testsuite/gcc.dg/wrapped-binop-simplify.c
> @@ -1,4 +1,4 @@
> -/* { dg-do compile } */
> +/* { dg-do compile { target { { i?86-*-* x86_64-*-* s390*-*-* } && lp64 } }
> } */
> /* { dg-options "-O2 -fdump-tree-vrp2-details" } */
> /* { dg-final { scan-tree-dump-times "gimple_simplified to" 4 "vrp2" } } */
Yes, I think so.
