Intermangled OpenMP 5.0 and 5.1.
OpenMP 5.1:
- 'capture' is no longer an atomic-clause
- can only be used with 'update' (which is implied if absent)
(a) The patch (accidentally) accepted the OpenMP 5 syntax.
→ now rejects 'capture' + 'update'
Additionally:
(b) There was a copy and paste error
On Mon, Nov 02, 2020 at 12:52:03PM +0100, Tobias Burnus wrote:
> Fortran: OpenMP - fixes for omp atomic [PR97655]
>
> gcc/fortran/ChangeLog:
>
> PR fortran/97655
> * openmp.c (gfc_match_omp_atomic): Fix mem-order handling;
> reject specifying update + capture together.
>
> gcc/
On Thu, Oct 29, 2020 at 06:05:41PM +0100, Tobias Burnus wrote:
> gcc/fortran/ChangeLog:
>
> * dump-parse-tree.c (show_omp_clauses): Handle atomic clauses.
> (show_omp_node): Call it for atomic.
> * gfortran.h (enum gfc_omp_atomic_op): Add GFC_OMP_ATOMIC_UNSET,
> remove GFC_
Forgot to add fortran@.
On 29.10.20 18:05, Tobias Burnus wrote:
The parser partially anticipates the upcoming OpenMP 5.1 changes, which
adds some more clauses - but otherwise does not update it for OpenMP 5.1,
yet. In particular, the "omp end atomic" for capture is still required
and
the memory-
The parser partially anticipates the upcoming OpenMP 5.1 changes, which
adds some more clauses - but otherwise does not update it for OpenMP 5.1,
yet. In particular, the "omp end atomic" for capture is still required and
the memory-order-clause restrictions still apply.
I am a bit unsure about ho
