2011/7/13 Richard Guenther <richard.guent...@gmail.com>:
> On Wed, Jul 13, 2011 at 9:33 AM, Kai Tietz <ktiet...@googlemail.com> wrote:
>> Hello,
>>
>> This patch adds support to fold_binary_loc for one-bit precision
>> typed bitwise-or expression.
>
> Seems to be a fallout of the missing TRUTH_NOT conversion as well.
>
>> ChangeLog
>>
>> 2011-07-13  Kai Tietz  <kti...@redhat.com>
>>
>>        * fold-const.c (fold_binary_loc): Add
>>        support for one-bit bitwise-or optimizeation.
>>
>> Bootstrapped and regression tested with prior patches of this series
>> for x86_64-pc-linux-gnu.
>> Ok for apply?
>>
>> Regards,
>> Kai
>>
>> Index: gcc/gcc/fold-const.c
>> ===================================================================
>> --- gcc.orig/gcc/fold-const.c   2011-07-13 08:23:29.000000000 +0200
>> +++ gcc/gcc/fold-const.c        2011-07-13 08:59:04.011620200 +0200
>> @@ -10688,6 +10688,52 @@ fold_binary_loc (location_t loc,
>>          return omit_one_operand_loc (loc, type, t1, arg0);
>>        }
>>
>> +      if (TYPE_PRECISION (type) == 1 && INTEGRAL_TYPE_P (type))
>> +        {
>> +         /* If arg0 is constant zero, drop it.  */
>> +         if (TREE_CODE (arg0) == INTEGER_CST && integer_zerop (arg0))
>> +           return non_lvalue_loc (loc, fold_convert_loc (loc, type, arg1));
>> +         if (TREE_CODE (arg0) == INTEGER_CST && ! integer_zerop (arg0))
>> +           return omit_one_operand_loc (loc, type, arg0, arg1);
>> +
>> +         /* !X | X is always true. ~X | X is always true.  */
>> +         if ((TREE_CODE (arg0) == TRUTH_NOT_EXPR
>> +              || TREE_CODE (arg0) == BIT_NOT_EXPR)
>> +             && operand_equal_p (TREE_OPERAND (arg0, 0), arg1, 0))
>> +           return omit_one_operand_loc (loc, type, integer_one_node, arg1);
>> +         /* X | !X is always true. X | ~X is always true.  */
>> +         if ((TREE_CODE (arg1) == TRUTH_NOT_EXPR
>> +             || TREE_CODE (arg1) == BIT_NOT_EXPR)
>> +             && operand_equal_p (arg0, TREE_OPERAND (arg1, 0), 0))
>> +           return omit_one_operand_loc (loc, type, integer_one_node, arg0);
>> +
>> +         /* (X & !Y) | (!X & Y) is X ^ Y */
>> +         if (TREE_CODE (arg0) == BIT_AND_EXPR
>> +             && TREE_CODE (arg1) == BIT_AND_EXPR)
>> +           {
>> +             tree a0, a1, l0, l1, n0, n1;
>> +
>> +             a0 = fold_convert_loc (loc, type, TREE_OPERAND (arg1, 0));
>> +             a1 = fold_convert_loc (loc, type, TREE_OPERAND (arg1, 1));
>> +
>> +             l0 = fold_convert_loc (loc, type, TREE_OPERAND (arg0, 0));
>> +             l1 = fold_convert_loc (loc, type, TREE_OPERAND (arg0, 1));
>> +
>> +             n0 = fold_build1_loc (loc, TRUTH_NOT_EXPR, type, l0);
>> +             n1 = fold_build1_loc (loc, TRUTH_NOT_EXPR, type, l1);
>> +
>> +             if ((operand_equal_p (n0, a0, 0)
>> +                  && operand_equal_p (n1, a1, 0))
>> +                 || (operand_equal_p (n0, a1, 0)
>> +                     && operand_equal_p (n1, a0, 0)))
>> +               return fold_build2_loc (loc, BIT_XOR_EXPR, type, l0, n1);
>> +           }
>> +
>> +         tem = fold_truth_andor (loc, code, type, arg0, arg1, op0, op1);
>> +         if (tem)
>> +           return tem;
>> +        }
>> +
>>       /* Canonicalize (X & C1) | C2.  */
>>       if (TREE_CODE (arg0) == BIT_AND_EXPR
>>          && TREE_CODE (arg1) == INTEGER_CST

Well, I wouldn't call it fallout.  As by this we are able to handle
things like ~(X >= B) and see that it can be converted to X < B.  The
point here is that we avoid that fold re-introduces here the TRUTH
variants for the bitwise ones (for sure some parts are redudant and
might be something to be factored out like we did for truth_andor
function). Also we catch by this patterns like ~X op ~Y and convert
them to ~(X op Y), which is just valid for one-bit precision typed X
and Y.
As in general !x is not the same as ~x, beside x has one-bit precision
integeral type.

 I will adjust patches so, that for one-bit precision type we alway
use here instead BIT_NOT_EXPR (instead of TRUTH_NOT). This is
reasonable.

Reply via email to