genmatch currently has a difficulty to decide whether a duplicate
structural match is really duplicate as uses of captures within
predicates or in C code can be order dependent. For example
a reported duplicate results in
{
tree captures[4] ATTRIBUTE_UNUSED = { _p1, _p0, _q20, _q21 }
if (gimple_simplify_112 (res_op, seq, valueize, type, captures))
return true;
}
{
tree captures[4] ATTRIBUTE_UNUSED = { _p1, _p0, _q21, _q20 };
if (gimple_simplify_112 (res_op, seq, valueize, type, captures))
return true;
}
where the difference is only in _q20 and _q21 being swapped but
that resulting in a call to bitwise_inverted_equal_p (_p1, X)
with X once _q20 and once _q21. That is, we treat bare
captures as equal for reporting duplicates.
Due to bitwise_inverted_equal_p there are meanwhile a _lot_ of
duplicates reported that are not actual duplicates.
The following removes some that are though, as the operands are
only passed to types_match.
Bootstrapped and tested on x86_64-unknown-linux-gnu.
* match.pd (.SAT_ADD patterns using IFN_ADD_OVERFLOW): Remove :c that
only causes duplicate patterns.
---
gcc/match.pd | 10 +++++-----
1 file changed, 5 insertions(+), 5 deletions(-)
diff --git a/gcc/match.pd b/gcc/match.pd
index fd1d8bcc776..656eb0e6987 100644
--- a/gcc/match.pd
+++ b/gcc/match.pd
@@ -3146,30 +3146,30 @@ DEFINE_INT_AND_FLOAT_ROUND_FN (RINT)
(match (unsigned_integer_sat_add @0 @1)
/* SUM = ADD_OVERFLOW (X, Y)
SAT_U_ADD = REALPART (SUM) | -IMAGPART (SUM) */
- (bit_ior:c (realpart (IFN_ADD_OVERFLOW:c@2 @0 @1)) (negate (imagpart @2)))
+ (bit_ior:c (realpart (IFN_ADD_OVERFLOW@2 @0 @1)) (negate (imagpart @2)))
(if (types_match (type, @0, @1))))
(match (unsigned_integer_sat_add @0 @1)
/* SUM = ADD_OVERFLOW (X, Y)
SAT_U_ADD = REALPART (SUM) | -(IMAGPART (SUM) != 0) */
- (bit_ior:c (realpart (IFN_ADD_OVERFLOW:c@2 @0 @1))
+ (bit_ior:c (realpart (IFN_ADD_OVERFLOW@2 @0 @1))
(negate (convert (ne (imagpart @2) integer_zerop))))
(if (types_match (type, @0, @1))))
(match (unsigned_integer_sat_add @0 @1)
/* SUM = ADD_OVERFLOW (X, Y)
SAT_U_ADD = IMAGPART (SUM) == 0 ? REALPART (SUM) : -1 */
- (cond^ (eq (imagpart (IFN_ADD_OVERFLOW:c@2 @0 @1)) integer_zerop)
+ (cond^ (eq (imagpart (IFN_ADD_OVERFLOW@2 @0 @1)) integer_zerop)
(realpart @2) integer_minus_onep)
(if (types_match (type, @0, @1))))
(match (unsigned_integer_sat_add @0 @1)
/* SUM = ADD_OVERFLOW (X, Y)
SAT_U_ADD = IMAGPART (SUM) != 0 ? -1 : REALPART (SUM) */
- (cond^ (ne (imagpart (IFN_ADD_OVERFLOW:c@2 @0 @1)) integer_zerop)
+ (cond^ (ne (imagpart (IFN_ADD_OVERFLOW@2 @0 @1)) integer_zerop)
integer_minus_onep (realpart @2))
(if (types_match (type, @0, @1))))
(match (unsigned_integer_sat_add @0 @1)
/* SUM = ADD_OVERFLOW (X, IMM)
SAT_U_ADD = IMAGPART (SUM) != 0 ? -1 : REALPART (SUM) */
- (cond^ (ne (imagpart (IFN_ADD_OVERFLOW:c@2 @0 INTEGER_CST@1)) integer_zerop)
+ (cond^ (ne (imagpart (IFN_ADD_OVERFLOW@2 @0 INTEGER_CST@1)) integer_zerop)
integer_minus_onep (realpart @2))
(if (types_match (type, @0) && int_fits_type_p (@1, type)))))
--
2.43.0
