The following adjusts vectorizer code generation to avoid splitting
out address increments for invariant addresses which causes objects
to get TREE_ADDRESSABLE when not necessary.
Bootstrapped and tested on x86_64-unknown-linux-gnu, pushed.
* tree-vect-data-refs.cc (bump_vector_ptr): Return an
invariant updated address when the input was invariant.
---
gcc/tree-vect-data-refs.cc | 8 ++++++++
1 file changed, 8 insertions(+)
diff --git a/gcc/tree-vect-data-refs.cc b/gcc/tree-vect-data-refs.cc
index 609cacc4971..b279a82551e 100644
--- a/gcc/tree-vect-data-refs.cc
+++ b/gcc/tree-vect-data-refs.cc
@@ -5195,6 +5195,14 @@ bump_vector_ptr (vec_info *vinfo,
if (TREE_CODE (dataref_ptr) == SSA_NAME)
new_dataref_ptr = copy_ssa_name (dataref_ptr);
+ else if (is_gimple_min_invariant (dataref_ptr))
+ /* When possible avoid emitting a separate increment stmt that will
+ force the addressed object addressable. */
+ return build1 (ADDR_EXPR, TREE_TYPE (dataref_ptr),
+ fold_build2 (MEM_REF,
+ TREE_TYPE (TREE_TYPE (dataref_ptr)),
+ dataref_ptr,
+ fold_convert (ptr_type_node, update)));
else
new_dataref_ptr = make_ssa_name (TREE_TYPE (dataref_ptr));
incr_stmt = gimple_build_assign (new_dataref_ptr, POINTER_PLUS_EXPR,
--
2.35.3
