diff options
| author | Yannick Moy <moy@adacore.com> | 2019-09-17 08:02:56 +0000 |
|---|---|---|
| committer | Pierre-Marie de Rodat <pmderodat@gcc.gnu.org> | 2019-09-17 08:02:56 +0000 |
| commit | e34716b8dd8836a565b4cf0c26f7244161f194f1 (patch) | |
| tree | 65a5f634701b28725bcded695f9b90b8a2b74351 /gcc/fortran/expr.c | |
| parent | 0d4fcc9f622ee428849091c51108a4823819e9a6 (diff) | |
| download | gcc-e34716b8dd8836a565b4cf0c26f7244161f194f1.zip gcc-e34716b8dd8836a565b4cf0c26f7244161f194f1.tar.gz gcc-e34716b8dd8836a565b4cf0c26f7244161f194f1.tar.bz2 | |
[Ada] Fix rounding of fixed-point arithmetic operation
Fixed-point multiplication, division and conversion may lead to calling
the function Double_Divide in s-arit64 runtime unit. In the special case
where arguments have the special values X = -2**63 and the absolute
value of the product of its other arguments Y*Z = 2**64, the rounded
value should be either -1 or 1, but currently Double_Divide returns a
quotient of 0.
Rounding only applies when Round attribute is called on the arithmetic
operation for a decimal fixed-point result, or the result type is
integer.
This fixes correctly applies rounding away from 0 in that special case.
2019-09-17 Yannick Moy <moy@adacore.com>
gcc/ada/
* libgnat/s-arit64.adb (Double_Divide): Correctly handle the
special case when rounding.
gcc/testsuite/
* gnat.dg/fixedpnt7.adb: New testcase.
From-SVN: r275796
Diffstat (limited to 'gcc/fortran/expr.c')
0 files changed, 0 insertions, 0 deletions