https://gcc.gnu.org/bugzilla/show_bug.cgi?id=71598
Richard Biener <rguenth at gcc dot gnu.org> changed:
What |Removed |Added
----------------------------------------------------------------------------
CC| |ebotcazou at gcc dot gnu.org,
| |jason at gcc dot gnu.org,
| |jsm28 at gcc dot gnu.org
--- Comment #3 from Richard Biener <rguenth at gcc dot gnu.org> ---
get_alias_set does nothing special for ENUMERAL_TYPE which means distinct
enums get distinct alias sets. Iff the C standard says that from a TBAA
perspective an enumeral type is to be treated as one of the standard
integer types then we need a way to represent such type in the middle-end
or the frontends need to not expose the distinct ENUMERAL_TYPE (at least
in accesses) but whatever underlying type is to be used.
Note since this targets the middle-end behavior the required language
semantics of other languages play a role as well.
A conservative middle-end only patch could look like the following - possibly
also a good idea for LTO. Eventually we could settle on the ENUMERAL_TYPEs
TYPE_VALUES first entry type to specify that underlying type, but that would
need agreement. It looks like TREE_TYPE of an ENUMERAL_TYPE is also not
"taken" (but all FEs needing the get_alias_set treatment would need adjustment,
at least it could be conditional on TREE_TYPE set). The TYPE_VALUES idea
would leave us with no way to handle empty enums. Another thing to look
at would be TREE_TYPE (TYPE_MIN/MAX_VALUE), for the testcase and the C FE
it is unsigned int.
diff --git a/gcc/alias.c b/gcc/alias.c
index b64e3ea264d..8fbf34bf4dc 100644
--- a/gcc/alias.c
+++ b/gcc/alias.c
@@ -931,6 +931,11 @@ get_alias_set (tree t)
if (set != -1)
return set;
+ /* Enums inter-operate with the corresponding integer type. */
+ else if (TREE_CODE (t) == ENUMERAL_TYPE)
+ return get_alias_set (build_nonstandard_integer_type
+ (TYPE_PRECISION (t), TYPE_UNSIGNED (t)));
+
/* There are no objects of FUNCTION_TYPE, so there's no point in
using up an alias set for them. (There are, of course, pointers
and references to functions, but that's different.) */
