https://gcc.gnu.org/bugzilla/show_bug.cgi?id=107346
Eric Botcazou <ebotcazou at gcc dot gnu.org> changed:
What |Removed |Added
----------------------------------------------------------------------------
CC| |ebotcazou at gcc dot gnu.org
--- Comment #8 from Eric Botcazou <ebotcazou at gcc dot gnu.org> ---
> I am wondering whether I should try to support this, or bail out of
> vect_check_gather_scatter if pbitpos is not a multiple of BITS_PER_UNIT. The
> latter obviously feels safer.
It turns out that this vect_check_gather_scatter problem was latent since I can
reproduce it with the GCC 12 compiler in Ada, and the following fix:
diff --git a/gcc/tree-vect-data-refs.cc b/gcc/tree-vect-data-refs.cc
index e32b1779c5d..a22eb2d763d 100644
--- a/gcc/tree-vect-data-refs.cc
+++ b/gcc/tree-vect-data-refs.cc
@@ -4013,7 +4013,7 @@ vect_check_gather_scatter (stmt_vec_info stmt_info,
loop_vec_info loop_vinfo,
that can be gimplified before the loop. */
base = get_inner_reference (base, &pbitsize, &pbitpos, &off, &pmode,
&punsignedp, &reversep, &pvolatilep);
- if (reversep)
+ if (!multiple_p (pbitpos, BITS_PER_UNIT) || reversep)
return false;
poly_int64 pbytepos = exact_div (pbitpos, BITS_PER_UNIT);
is sufficient to get rid of it.
