When fixing PR109672 I noticed we let SImode AND through when
target_support_p even though it isn't word_mode and I didn't want to
change that but had to catch the case where SImode PLUS is supported
but emulated vectors rely on it being word_mode. The following
makes sure to preserve the word_mode check when !target_support_p
to avoid excessive lowering later even for bit operations.
Bootstrapped and tested on x86_64-unknown-linux-gnu, pushed.
PR tree-optimization/109721
* tree-vect-stmts.cc (vectorizable_operation): Make sure
to test word_mode for all !target_support_p operations.
---
gcc/tree-vect-stmts.cc | 10 ++++++++--
1 file changed, 8 insertions(+), 2 deletions(-)
diff --git a/gcc/tree-vect-stmts.cc b/gcc/tree-vect-stmts.cc
index 3ad6a7d28d7..cf5194ea444 100644
--- a/gcc/tree-vect-stmts.cc
+++ b/gcc/tree-vect-stmts.cc
@@ -6389,9 +6389,15 @@ vectorizable_operation (vec_info *vinfo,
if (dump_enabled_p ())
dump_printf_loc (MSG_MISSED_OPTIMIZATION, vect_location,
"op not supported by target.\n");
- /* Check only during analysis. */
- if (((code == PLUS_EXPR || code == MINUS_EXPR || code == NEGATE_EXPR)
+ /* When vec_mode is not a vector mode and we verified ops we
+ do not have to lower like AND are natively supported let
+ those through even when the mode isn't word_mode. For
+ ops we have to lower the lowering code assumes we are
+ dealing with word_mode. */
+ if ((((code == PLUS_EXPR || code == MINUS_EXPR || code == NEGATE_EXPR)
+ || !target_support_p)
&& maybe_ne (GET_MODE_SIZE (vec_mode), UNITS_PER_WORD))
+ /* Check only during analysis. */
|| (!vec_stmt && !vect_can_vectorize_without_simd_p (code)))
{
if (dump_enabled_p ())
--
2.35.3
