The pattern `(a == b) | ((a|b) != 0)` uses build_one_cst to build boolean true
but boolean can be a signed multi-bit type. So this changes the result to
use constant_boolean_node isntead.
`(a != b) & ((a|b) == 0)` has a similar issue but in that case it is less likely
to be an issue as false is almost always just 0 but this changes it to be consistent.
Pushed as obvious after a bootstrap/test on x86_64-linux-gnu.
PR tree-optimization/122296
gcc/ChangeLog:
* match.pd (`(a == b) | ((a|b) != 0)`): Fix true value.
gcc/testsuite/ChangeLog:
* gcc.dg/tree-ssa/int-bwise-opt-vect01.c: New test.
Signed-off-by: Andrew Pinski <andrew.pinski@oss.qualcomm.com>
(eq @0 @1))
(simplify
(bit_and:c (ne:c @0 @1) (eq (bit_ior @0 @1) integer_zerop))
- { build_zero_cst (type); })
+ { constant_boolean_node (false, type); })
(simplify
(bit_ior:c (eq:c @0 @1) (ne (bit_ior @0 @1) integer_zerop))
- { build_one_cst (type); })
+ { constant_boolean_node (true, type); })
#endif
/* These was part of minmax phiopt. */
--- /dev/null
+/* { dg-do compile } */
+/* { dg-options "-O1 -fdump-tree-forwprop1 -fdump-tree-optimized" } */
+
+/* PR tree-optimization/122296 */
+
+typedef unsigned type1 __attribute__((vector_size(sizeof(unsigned))));
+
+type1 f(type1 a, type1 b)
+{
+ type1 c = a == b;
+ type1 d = (a|b) != 0;
+ return c | d;
+}
+
+/* { dg-final { scan-tree-dump-not "VEC_COND_EXPR " "optimized" } } */
+/* { dg-final { scan-tree-dump-not "VEC_COND_EXPR " "forwprop1" } } */
+