On Thu, May 7, 2020 at 9:48 AM Jakub Jelinek via Gcc-patches
<gcc-patches@gcc.gnu.org> wrote:
>
> Hi!
>
> If the second argument of __builtin_speculation_safe_value is
> error_mark_node (or has such a type), we ICE during
> useless_typ_conversion_p.
>
> Fixed thusly, bootstrapped/regtested on x86_64-linux and i686-linux, ok for
> trunk?
OK.
Richard.
> 202-05-06 Jakub Jelinek <ja...@redhat.com>
>
> PR c/94968
> * c-common.c (speculation_safe_value_resolve_params): Return false if
> error_operand_p (val2).
> (resolve_overloaded_builtin) <case BUILT_IN_SPECULATION_SAFE_VALUE_N>:
> Remove extraneous semicolon.
>
> * gcc.dg/pr94968.c: New test.
>
> --- gcc/c-family/c-common.c.jj 2020-04-27 16:10:23.795726234 +0200
> +++ gcc/c-family/c-common.c 2020-05-06 12:37:40.708539995 +0200
> @@ -6716,6 +6716,8 @@ speculation_safe_value_resolve_params (l
> tree val2 = (*params)[1];
> if (TREE_CODE (TREE_TYPE (val2)) == ARRAY_TYPE)
> val2 = default_conversion (val2);
> + if (error_operand_p (val2))
> + return false;
> if (!(TREE_TYPE (val) == TREE_TYPE (val2)
> || useless_type_conversion_p (TREE_TYPE (val), TREE_TYPE (val2))))
> {
> @@ -7400,7 +7402,7 @@ resolve_overloaded_builtin (location_t l
> {
> tree new_function, first_param, result;
> enum built_in_function fncode
> - = speculation_safe_value_resolve_call (function, params);;
> + = speculation_safe_value_resolve_call (function, params);
>
> if (fncode == BUILT_IN_NONE)
> return error_mark_node;
> --- gcc/testsuite/gcc.dg/pr94968.c.jj 2020-05-06 12:50:10.865191796 +0200
> +++ gcc/testsuite/gcc.dg/pr94968.c 2020-05-06 12:49:52.366471532 +0200
> @@ -0,0 +1,8 @@
> +/* PR c/94968 */
> +/* { dg-do compile } */
> +
> +int
> +foo (void)
> +{
> + __builtin_speculation_safe_value (1, x); /* { dg-error "undeclared" }
> */
> +} /* { dg-message "each
> undeclared identifier is reported only once" "" { target *-*-* } .-1 } */
>
> Jakub
>
