This is the mail archive of the
fortran@gcc.gnu.org
mailing list for the GNU Fortran project.
Re: Problem with threadprivate procedure pointers with gfortran 4.6
- From: Jakub Jelinek <jakub at redhat dot com>
- To: Andrew Benson <abenson at its dot caltech dot edu>
- Cc: fortran at gcc dot gnu dot org
- Date: Fri, 26 Aug 2011 12:15:46 +0200
- Subject: Re: Problem with threadprivate procedure pointers with gfortran 4.6
- References: <201108252025.42064.abenson@its.caltech.edu>
- Reply-to: Jakub Jelinek <jakub at redhat dot com>
On Thu, Aug 25, 2011 at 08:25:41PM -0700, Andrew Benson wrote:
> I'm finding what appears to be incorrect behavior when using threadprivate
> procedure pointers with gfortran 4.6. The problem is reminiscent of a previous
> problem that I had with gfortran 4.4:
>
> http://gcc.gnu.org/ml/fortran/2010-02/msg00080.html
>
> That problem was never tracked down because it didn't occur in 4.5. The new
> problem is similar, but different. I've modified the test case that Janus
> created for the previous problem:
Ugh, you've been told a year and half ago that what you are trying to do
is undefined behavior, OpenMP standard, including 3.1, doesn't support
Fortran 2003 nor Fortran 2008, but you keep ignoring that. In 4.5 it wasn't
any different, globalProc wasn't TLS either.
That said, I'm testing a patch to, as an extension, handle procedure
pointers for gfortran 4.7. Your testcase can't work on > 2 CPU system
even if globalProc is TLS (except if OMP_NUM_THREADS=2 (or =1)).
Jakub