The following fixes the mask argument generation for SIMD clone
calls under either loop masking or when the actual call is not
masked but only a inbranch simd clone is available. The issue
was that we tried to directly convert the vector mask to the
call argument type but SIMD clone masks require 1 or 0 (which
could be even float) values for mask elements so we have to
resort to a VEC_COND_EXPR to generate them just like we do for
regular passing of the mask.
Bootstrapped and tested on x86_64-unknown-linux-gnu, pushed.
PR tree-optimization/112405
* tree-vect-stmts.cc (vectorizable_simd_clone_call):
Properly handle invariant and/or loop mask passing.
---
gcc/tree-vect-stmts.cc | 62 +++++++++++++++++++++++-------------------
1 file changed, 34 insertions(+), 28 deletions(-)
diff --git a/gcc/tree-vect-stmts.cc b/gcc/tree-vect-stmts.cc
index eefb1eec1ef..65883e04ad7 100644
--- a/gcc/tree-vect-stmts.cc
+++ b/gcc/tree-vect-stmts.cc
@@ -4814,36 +4814,42 @@ vectorizable_simd_clone_call (vec_info *vinfo,
stmt_vec_info stmt_info,
else
mask = vect_build_all_ones_mask (vinfo, stmt_info, masktype);
- if (!useless_type_conversion_p (TREE_TYPE (mask), masktype))
+ gassign *new_stmt;
+ if (SCALAR_INT_MODE_P (bestn->simdclone->mask_mode))
{
- gassign *new_stmt;
- if (bestn->simdclone->mask_mode != VOIDmode)
- {
- /* This means we are dealing with integer mask modes.
- First convert to an integer type with the same size as
- the current vector type. */
- unsigned HOST_WIDE_INT intermediate_size
- = tree_to_uhwi (TYPE_SIZE (TREE_TYPE (mask)));
- tree mid_int_type =
- build_nonstandard_integer_type (intermediate_size, 1);
- mask = build1 (VIEW_CONVERT_EXPR, mid_int_type, mask);
- new_stmt
- = gimple_build_assign (make_ssa_name (mid_int_type),
- mask);
- gsi_insert_before (gsi, new_stmt, GSI_SAME_STMT);
- /* Then zero-extend to the mask mode. */
- mask = fold_build1 (NOP_EXPR, masktype,
- gimple_get_lhs (new_stmt));
- }
- else
- mask = build1 (VIEW_CONVERT_EXPR, masktype, mask);
-
- new_stmt = gimple_build_assign (make_ssa_name (masktype),
- mask);
- vect_finish_stmt_generation (vinfo, stmt_info,
- new_stmt, gsi);
- mask = gimple_assign_lhs (new_stmt);
+ /* This means we are dealing with integer mask modes.
+ First convert to an integer type with the same size as
+ the current vector type. */
+ unsigned HOST_WIDE_INT intermediate_size
+ = tree_to_uhwi (TYPE_SIZE (TREE_TYPE (mask)));
+ tree mid_int_type =
+ build_nonstandard_integer_type (intermediate_size, 1);
+ mask = build1 (VIEW_CONVERT_EXPR, mid_int_type, mask);
+ new_stmt
+ = gimple_build_assign (make_ssa_name (mid_int_type),
+ mask);
+ gsi_insert_before (gsi, new_stmt, GSI_SAME_STMT);
+ /* Then zero-extend to the mask mode. */
+ mask = fold_build1 (NOP_EXPR, masktype,
+ gimple_get_lhs (new_stmt));
+ }
+ else if (bestn->simdclone->mask_mode == VOIDmode)
+ {
+ tree one = fold_convert (TREE_TYPE (masktype),
+ integer_one_node);
+ tree zero = fold_convert (TREE_TYPE (masktype),
+ integer_zero_node);
+ mask = build3 (VEC_COND_EXPR, masktype, mask,
+ build_vector_from_val (masktype, one),
+ build_vector_from_val (masktype, zero));
}
+ else
+ gcc_unreachable ();
+
+ new_stmt = gimple_build_assign (make_ssa_name (masktype), mask);
+ vect_finish_stmt_generation (vinfo, stmt_info,
+ new_stmt, gsi);
+ mask = gimple_assign_lhs (new_stmt);
vargs.safe_push (mask);
}
}
--
2.35.3
