When backporting the PR109473 fix I failed to realize its testcase
will run into an unrelated similar bug. With GCC 12 the code has
seen substantial refactoring so the following applies a local fix
to make sure we are using the correct types when building initial
values for reductions.
Bootstrapped and tested on x86_64-unknown-linux-gnu, pushed.
PR tree-optimization/109473
* tree-vect-loop.cc (get_initial_def_for_reduction):
Convert the scalar values to the vector component type
before using it to build the vector for the initial value.
---
gcc/tree-vect-loop.c | 8 ++++++--
1 file changed, 6 insertions(+), 2 deletions(-)
diff --git a/gcc/tree-vect-loop.c b/gcc/tree-vect-loop.c
index bb021ccf382..ebdba65c55e 100644
--- a/gcc/tree-vect-loop.c
+++ b/gcc/tree-vect-loop.c
@@ -4730,10 +4730,14 @@ get_initial_def_for_reduction (loop_vec_info loop_vinfo,
else
def_for_init = build_int_cst (scalar_type, int_init_val);
- if (adjustment_def || operand_equal_p (def_for_init, init_val, 0))
+ bool same_p = operand_equal_p (def_for_init, init_val, 0);
+ init_val = gimple_convert (&stmts, TREE_TYPE (vectype), init_val);
+ def_for_init = gimple_convert (&stmts, TREE_TYPE (vectype),
def_for_init);
+
+ if (adjustment_def || same_p)
{
/* Option1: the first element is '0' or '1' as well. */
- if (!operand_equal_p (def_for_init, init_val, 0))
+ if (!same_p)
*adjustment_def = init_val;
init_def = gimple_build_vector_from_val (&stmts, vectype,
def_for_init);
--
2.35.3
