match.pd: Demote IFN_{ADD,SUB,MUL}_OVERFLOW operands [PR99591]

The overflow builtins work on infinite precision integers and then convert
to the result type's precision, so any argument promotions are useless.
The expand_arith_overflow expansion is able to demote the arguments itself
through get_range_pos_neg and get_min_precision calls and if needed promote
to whatever mode it decides to perform the operations in, but if there are
any promotions it demoted, those are already expanded.  Normally combine
would remove the useless sign or zero extensions when it sees the result
of those is only used in a lowpart subreg, but typically those lowpart
subregs appear multiple times in the pattern so that they describe properly
the overflow behavior and combine gives up, so we end up with e.g.
        movswl  %si, %esi
        movswl  %di, %edi
        imulw   %si, %di
        seto    %al
where both movswl insns are useless.

The following patch fixes it by demoting operands of the ifns (only gets
rid of integral to integral conversions that increase precision).
While IFN_{ADD,MUL}_OVERFLOW are commutative and just one simplify would be
enough, IFN_SUB_OVERFLOW is not, therefore two simplifications.

2021-09-02  Jakub Jelinek  <jakub@redhat.com>

	PR tree-optimization/99591
	* match.pd: Demote operands of IFN_{ADD,SUB,MUL}_OVERFLOW if they
	were promoted.

	* gcc.target/i386/pr99591.c: New test.
	* gcc.target/i386/pr97950.c: Match or reject setb or jn?b instructions
	together with seta or jn?a.

diff --git a/gcc/match.pd b/gcc/match.pd
index f421c74b62c77e2708448bc7951faf84a9fbe093..f920bc4b7c1c3b4d7de71eb91592f14026043959 100644 (file)
--- a/gcc/match.pd
+++ b/gcc/match.pd
@@ -5587,6 +5587,21 @@ DEFINE_INT_AND_FLOAT_ROUND_FN (RINT)
    (with { tree t = TREE_TYPE (@0), cpx = build_complex_type (t); }
     (cmp (imagpart (IFN_MUL_OVERFLOW:cpx @0 @1)) { build_zero_cst (t); })))))
 
+/* Demote operands of IFN_{ADD,SUB,MUL}_OVERFLOW.  */
+(for ovf (IFN_ADD_OVERFLOW IFN_SUB_OVERFLOW IFN_MUL_OVERFLOW)
+ (simplify
+  (ovf (convert@2 @0) @1)
+  (if (INTEGRAL_TYPE_P (TREE_TYPE (@0))
+       && INTEGRAL_TYPE_P (TREE_TYPE (@2))
+       && TYPE_PRECISION (TREE_TYPE (@2)) > TYPE_PRECISION (TREE_TYPE (@0)))
+   (ovf @0 @1)))
+ (simplify
+  (ovf @1 (convert@2 @0))
+  (if (INTEGRAL_TYPE_P (TREE_TYPE (@0))
+       && INTEGRAL_TYPE_P (TREE_TYPE (@2))
+       && TYPE_PRECISION (TREE_TYPE (@2)) > TYPE_PRECISION (TREE_TYPE (@0)))
+   (ovf @1 @0))))
+
 /* Simplification of math builtins.  These rules must all be optimizations
    as well as IL simplifications.  If there is a possibility that the new
    form could be a pessimization, the rule should go in the canonicalization

diff --git a/gcc/testsuite/gcc.target/i386/pr97950.c b/gcc/testsuite/gcc.target/i386/pr97950.c
index 277311d9a4ea1db3d97133392231c6e02a7586a7..185d73378cb3987328edc36b17319b18d331c4ae 100644 (file)
--- a/gcc/testsuite/gcc.target/i386/pr97950.c
+++ b/gcc/testsuite/gcc.target/i386/pr97950.c
@@ -1,10 +1,10 @@
 /* PR target/95950 */
 /* { dg-do compile } */
 /* { dg-options "-O2 -mtune=generic" } */
-/* { dg-final { scan-assembler-times "\tseta\t" 4 } } */
+/* { dg-final { scan-assembler-times "\tset\[ab]\t" 4 } } */
 /* { dg-final { scan-assembler-times "\tseto\t" 16 } } */
 /* { dg-final { scan-assembler-times "\tsetc\t" 4 } } */
-/* { dg-final { scan-assembler-not "\tjn?a\t" } } */
+/* { dg-final { scan-assembler-not "\tjn?\[ab]\t" } } */
 /* { dg-final { scan-assembler-not "\tjn?o\t" } } */
 /* { dg-final { scan-assembler-not "\tjn?c\t" } } */
 

diff --git a/gcc/testsuite/gcc.target/i386/pr99591.c b/gcc/testsuite/gcc.target/i386/pr99591.c
new file mode 100644 (file)
index 0000000..01d8485
--- /dev/null
+++ b/gcc/testsuite/gcc.target/i386/pr99591.c
@@ -0,0 +1,32 @@
+/* PR tree-optimization/99591 */
+/* { dg-do compile } */
+/* { dg-options "-O2" } */
+/* { dg-final { scan-assembler-not "\tmovs\[bw]l\t" } } */
+
+int
+foo (signed char a, signed char b)
+{
+  signed char r;
+  return __builtin_add_overflow (a, b, &r);
+}
+
+int
+bar (short a, short b)
+{
+  short r;
+  return __builtin_add_overflow (a, b, &r);
+}
+
+int
+baz (signed char a, signed char b)
+{
+  signed char r;
+  return __builtin_add_overflow ((int) a, (int) b, &r);
+}
+
+int
+qux (short a, short b)
+{
+  short r;
+  return __builtin_add_overflow ((int) a, (int) b, &r);
+}