On Tue, Apr 24, 2012 at 9:10 AM, Jakub Jelinek <ja...@redhat.com> wrote:
> On Tue, Apr 24, 2012 at 08:05:35AM +0200, Richard Guenther wrote:
>> On Tue, Apr 24, 2012 at 1:01 AM, Jakub Jelinek <ja...@redhat.com> wrote:
>> I don't like using TYPE_MAX_VALUE in VRP - at least please use vrp_val_max.
>> For enums this can be not what you expect(?)
>>
>> Please consider adding a double_int_max_value (unsigned prec, bool sign)
>> and double_int_min_value.
>
> So like this instead?
Yes!
Thanks,
Richard.
> 2012-04-24 Jakub Jelinek <ja...@redhat.com>
>
> PR tree-optimization/53058
> * double-int.h (double_int_max_value, double_int_min_value): New
> functions.
> * tree-vrp.c (register_edge_assert_for_2): Compare mask
> for LE_EXPR or GT_EXPR with double_int_max_value
> instead of double_int_mask.
>
> * gcc.c-torture/compile/pr53058.c: New test.
>
> --- gcc/double-int.h.jj 2012-04-19 11:09:13.000000000 +0200
> +++ gcc/double-int.h 2012-04-24 08:42:42.655756406 +0200
> @@ -1,5 +1,5 @@
> /* Operations with long integers.
> - Copyright (C) 2006, 2007, 2008, 2010 Free Software Foundation, Inc.
> + Copyright (C) 2006, 2007, 2008, 2010, 2012 Free Software Foundation, Inc.
>
> This file is part of GCC.
>
> @@ -280,6 +280,26 @@ double_int_equal_p (double_int cst1, dou
> return cst1.low == cst2.low && cst1.high == cst2.high;
> }
>
> +/* Returns a maximum value for signed or unsigned integer
> + of precision PREC. */
> +
> +static inline double_int
> +double_int_max_value (unsigned int prec, bool uns)
> +{
> + return double_int_mask (prec - (uns ? 0 : 1));
> +}
> +
> +/* Returns a minimum value for signed or unsigned integer
> + of precision PREC. */
> +
> +static inline double_int
> +double_int_min_value (unsigned int prec, bool uns)
> +{
> + if (uns)
> + return double_int_zero;
> + return double_int_lshift (double_int_one, prec - 1, prec, false);
> +}
> +
>
> /* Legacy interface with decomposed high/low parts. */
>
> --- gcc/tree-vrp.c.jj 2012-04-23 11:11:21.000000000 +0200
> +++ gcc/tree-vrp.c 2012-04-24 09:06:29.116624564 +0200
> @@ -4565,6 +4565,7 @@ register_edge_assert_for_2 (tree name, e
> && INTEGRAL_TYPE_P (TREE_TYPE (name2))
> && IN_RANGE (tree_low_cst (cst2, 1), 1, prec - 1)
> && prec <= 2 * HOST_BITS_PER_WIDE_INT
> + && prec == GET_MODE_PRECISION (TYPE_MODE (TREE_TYPE (val)))
> && live_on_edge (e, name2)
> && !has_single_use (name2))
> {
> @@ -4598,8 +4599,10 @@ register_edge_assert_for_2 (tree name, e
> new_val = val2;
> else
> {
> + double_int maxval
> + = double_int_max_value (prec, TYPE_UNSIGNED (TREE_TYPE
> (val)));
> mask = double_int_ior (tree_to_double_int (val2), mask);
> - if (double_int_minus_one_p (double_int_sext (mask, prec)))
> + if (double_int_equal_p (mask, maxval))
> new_val = NULL_TREE;
> else
> new_val = double_int_to_tree (TREE_TYPE (val2), mask);
> --- gcc/testsuite/gcc.c-torture/compile/pr53058.c.jj 2012-04-23
> 15:53:41.489982650 +0200
> +++ gcc/testsuite/gcc.c-torture/compile/pr53058.c 2012-04-23
> 15:53:13.000000000 +0200
> @@ -0,0 +1,12 @@
> +/* PR tree-optimization/53058 */
> +
> +int a, b, c;
> +
> +void
> +foo ()
> +{
> + c = b >> 16;
> + if (c > 32767)
> + c = 0;
> + a = b;
> +}
>
> Jakub
