On Fri, 9 Oct 2015, Jan Hubicka wrote:
> >
> > Index: expr.c
> > ===================================================================
> > --- expr.c (revision 228604)
> > +++ expr.c (working copy)
> > @@ -6703,7 +6704,7 @@ store_field (rtx target, HOST_WIDE_INT b
> > emit_group_store (temp_target, temp, TREE_TYPE (exp), size);
> > temp = temp_target;
> > }
> > - else if (mode == BLKmode)
> > + else if (GET_MODE (temp) == BLKmode)
> > {
> > /* Handle calls that return BLKmode values in registers. */
> > if (REG_P (temp) && TREE_CODE (exp) == CALL_EXPR)
>
> This patch passed the testing, so if FIELD_DECL of VOIDmode referring BLKmode
> type is a sane thing, I guess this is a right fix. I would say that however
> the type of FIELD_DECL should be compatible with the type of COMPONENT_REF and
> that should be added to the Gimple operand testing and fixed at Ada side?
I think a FIELD_DECL with VOIDmode is odd. And yes, the type of
the COMPONENT_REF should be that of the FIELD_DECL (or a variant type
of it as we share FIELD_DECLs for record variants). We do verify this
however:
static bool
verify_types_in_gimple_reference (tree expr, bool require_lvalue)
{
...
if (TREE_CODE (expr) == COMPONENT_REF
&& !useless_type_conversion_p (TREE_TYPE (expr),
TREE_TYPE (TREE_OPERAND (expr,
1))))
{
error ("type mismatch in component reference");
debug_generic_stmt (TREE_TYPE (expr));
debug_generic_stmt (TREE_TYPE (TREE_OPERAND (expr, 1)));
return true;
}
though if the type is an aggregate useless_type_conversion_p doesn't do
much now.
Richard.
> Honza
>
>
--
Richard Biener <rguent...@suse.de>
SUSE LINUX GmbH, GF: Felix Imendoerffer, Jane Smithard, Graham Norton, HRB
21284 (AG Nuernberg)
