[OpenMP, Patch] PR fortran/47886 fix ICE in OMP task's if() clause

H.J. Lu hjl.tools@gmail.com
Sun Feb 27 18:12:00 GMT 2011


On Sat, Feb 26, 2011 at 11:38 AM, Tobias Burnus <burnus@net-b.de> wrote:
> Jakub Jelinek wrote:
>>>
>>> Build and regtested on x86-64-linux.
>>> OK for the trunk? What about GCC 4.4 and 4.5?
>>
>> Oh, forgot about that, can you please add a testcase for this?
>> If you don't get to it, I'll do it on Monday.
>
> Thanks for the reminder! I even regtested with a test case - but I forget to
> include it in the patch ...
>
> Committed attached patch as Rev. 170521. I will backport after regtesting
> the branches.
>
> Tobias
>

I checked in this patch into 4.4 branch.


-- 
H.J.
---
Index: ChangeLog
===================================================================
--- ChangeLog	(revision 170544)
+++ ChangeLog	(working copy)
@@ -1,4 +1,8 @@
-2011-01-27  Tobias Burnus  <burnus@net-b.de>
+2011-02-27  H.J. Lu  <hongjiu.lu@intel.com>
+
+	* testsuite/libgomp.fortran/task3.f90: Remove duplicated code.
+
+2011-02-27  Tobias Burnus  <burnus@net-b.de>

 	Backport from mainline
 	2011-02-27  Jakub Jelinek  <jakub@redhat.com>
Index: testsuite/libgomp.fortran/task3.f90
===================================================================
--- testsuite/libgomp.fortran/task3.f90	(revision 170544)
+++ testsuite/libgomp.fortran/task3.f90	(working copy)
@@ -25,30 +25,3 @@ program F03_2_7_1d
       call abort
    end if
 end program F03_2_7_1d
-! { dg-do run }
-! { dg-options "-fopenmp" }
-!
-! PR fortran/47886
-!
-! Test case contributed by Bill Long
-
-!  derived from OpenMP test OMP3f/F03_2_7_1d.F90
-program F03_2_7_1d
-   use omp_lib
-   implicit none
-   integer, parameter :: NT = 4
-   integer :: sum = 0
-
-   call omp_set_num_threads(NT);
-
-   !$omp parallel
-   !$omp task if(omp_get_num_threads() > 0)
-   !$omp atomic
-      sum = sum + 1
-   !$omp end task
-   !$omp end parallel
-   if (sum /= NT) then
-      print *, "FAIL - sum == ", sum, " (expected ", NT, ")"
-      call abort
-   end if
-end program F03_2_7_1d



More information about the Fortran mailing list