match.pd: Disallow side-effects in GENERIC for non-COND_EXPR to COND_EXPR simplifications [PR93744]

As the following testcases show (the first one reported, last two
found by code inspection), we need to disallow side-effects
in simplifications that turn some unconditional expression into conditional
one.  From my little understanding of genmatch.c, it is able to
automatically disallow side effects if the same operand is used multiple
times in the match pattern, maybe if it is used multiple times in the
replacement pattern, and if it is used in conditional contexts in the match
pattern, could it be taught to handle this case too?  If yes, perhaps
just the first hunk could be usable for 8/9 backports (+ the testcases).

2020-02-15  Jakub Jelinek  <jakub@redhat.com>

	PR tree-optimization/93744
	* match.pd (((m1 >/</>=/<= m2) * d -> (m1 >/</>=/<= m2) ? d : 0,
	A - ((A - B) & -(C cmp D)) -> (C cmp D) ? B : A,
	A + ((B - A) & -(C cmp D)) -> (C cmp D) ? B : A): For GENERIC, make
	sure @2 in the first and @1 in the other patterns has no side-effects.

	* gcc.c-torture/execute/pr93744-1.c: New test.
	* gcc.c-torture/execute/pr93744-2.c: New test.
	* gcc.c-torture/execute/pr93744-3.c: New test.
This commit is contained in:
Jakub Jelinek 2020-02-15 12:53:44 +01:00
parent 55b00d14f4
commit 187dd955db
6 changed files with 77 additions and 3 deletions

View File

@ -1,3 +1,11 @@
2020-02-15 Jakub Jelinek <jakub@redhat.com>
PR tree-optimization/93744
* match.pd (((m1 >/</>=/<= m2) * d -> (m1 >/</>=/<= m2) ? d : 0,
A - ((A - B) & -(C cmp D)) -> (C cmp D) ? B : A,
A + ((B - A) & -(C cmp D)) -> (C cmp D) ? B : A): For GENERIC, make
sure @2 in the first and @1 in the other patterns has no side-effects.
2020-02-15 David Malcolm <dmalcolm@redhat.com>
Bernd Edlinger <bernd.edlinger@hotmail.de>

View File

@ -1472,7 +1472,8 @@ DEFINE_INT_AND_FLOAT_ROUND_FN (RINT)
(for cmp (gt lt ge le)
(simplify
(mult (convert (cmp @0 @1)) @2)
(cond (cmp @0 @1) @2 { build_zero_cst (type); })))
(if (GIMPLE || !TREE_SIDE_EFFECTS (@2))
(cond (cmp @0 @1) @2 { build_zero_cst (type); }))))
/* For integral types with undefined overflow and C != 0 fold
x * C EQ/NE y * C into x EQ/NE y. */
@ -2709,7 +2710,8 @@ DEFINE_INT_AND_FLOAT_ROUND_FN (RINT)
&& TREE_CODE (TREE_TYPE (@4)) != BOOLEAN_TYPE
&& INTEGRAL_TYPE_P (TREE_TYPE (@5))
&& (TYPE_PRECISION (TREE_TYPE (@4)) >= TYPE_PRECISION (type)
|| !TYPE_UNSIGNED (TREE_TYPE (@4))))
|| !TYPE_UNSIGNED (TREE_TYPE (@4)))
&& (GIMPLE || !TREE_SIDE_EFFECTS (@1)))
(cond (cmp @2 @3) @1 @0)))
(simplify
(plus:c @0 (bit_and:c (minus @1 @0)
@ -2719,7 +2721,8 @@ DEFINE_INT_AND_FLOAT_ROUND_FN (RINT)
&& TREE_CODE (TREE_TYPE (@4)) != BOOLEAN_TYPE
&& INTEGRAL_TYPE_P (TREE_TYPE (@5))
&& (TYPE_PRECISION (TREE_TYPE (@4)) >= TYPE_PRECISION (type)
|| !TYPE_UNSIGNED (TREE_TYPE (@4))))
|| !TYPE_UNSIGNED (TREE_TYPE (@4)))
&& (GIMPLE || !TREE_SIDE_EFFECTS (@1)))
(cond (cmp @2 @3) @1 @0))))
/* Simplifications of shift and rotates. */

View File

@ -1,3 +1,10 @@
2020-02-15 Jakub Jelinek <jakub@redhat.com>
PR tree-optimization/93744
* gcc.c-torture/execute/pr93744-1.c: New test.
* gcc.c-torture/execute/pr93744-2.c: New test.
* gcc.c-torture/execute/pr93744-3.c: New test.
2020-02-14 Jakub Jelinek <jakub@redhat.com>
PR c++/61414

View File

@ -0,0 +1,14 @@
/* PR tree-optimization/93744 */
typedef int I;
int
main ()
{
int a = 0;
I b = 0;
(a > 0) * (b |= 2);
if (b != 2)
__builtin_abort ();
return 0;
}

View File

@ -0,0 +1,21 @@
/* PR tree-optimization/93744 */
int w;
int
foo (int x, int y, int z)
{
int r = z - ((z - w++) & -(x < y));
return r;
}
int
main ()
{
w = 4;
if (foo (5, 7, 12) != 4 || w != 5)
__builtin_abort ();
if (foo (7, 5, 12) != 12 || w != 6)
__builtin_abort ();
return 0;
}

View File

@ -0,0 +1,21 @@
/* PR tree-optimization/93744 */
int w;
int
foo (int x, int y, int z)
{
int r = z + ((w++ - z) & -(x < y));
return r;
}
int
main ()
{
w = 4;
if (foo (5, 7, 12) != 4 || w != 5)
__builtin_abort ();
if (foo (7, 5, 12) != 12 || w != 6)
__builtin_abort ();
return 0;
}