On Wed, Mar 09, 2022 at 10:43:56PM +0800, Chung-Lin Tang wrote:
> when troubleshooting building/running SPEC HPC 2021 with GCC with OpenMP
> offloading,
> specifically 534.hpgmgfv_t, an issue encountered in the benchmark was:
> when the benchmark was initializing and creating its data environment on the
> GPU,
> it was trying to map array sections where the base-pointer is actually NULL:
> ...
> for (block=0;block<3;++block) {
> #pragma omp target enter data
> map(to:level->restriction[shape].blocks[block][:length])
> // level->restriction[shape].blocks[block] == NULL for some values of index
> 'block'
> ...
>
> The benchmark appears to be assuming that such NULL base-pointers would
> simply be
> silently ignored, and the program would just keep running.
I'd say that looks like a clear SPEC HPC 2021 bug, array sections are
similar to array elements and using even &ptr[0] or &ptr[10] when ptr is NULL
is UB
in C/C++.
Perhaps it doesn't hurt to discuss it on omp-lang, but I wouldn't change
anything on the compiler side until those discussions.
> 2022-03-09 Chung-Lin Tang <clt...@codesourcery.com>
>
> libgomp/ChangeLog:
>
> * target.c (gomp_attach_pointer): When pointer is NULL,
> return instead of calling gomp_fatal.
> diff --git a/libgomp/target.c b/libgomp/target.c
> index 9017458885e..0e8bbd83c20 100644
> --- a/libgomp/target.c
> +++ b/libgomp/target.c
> @@ -796,8 +796,8 @@ gomp_attach_pointer (struct gomp_device_descr *devicep,
>
> if ((void *) target == NULL)
> {
> - gomp_mutex_unlock (&devicep->lock);
> - gomp_fatal ("attempt to attach null pointer");
> + n->aux->attach_count[idx] = 0;
> + return;
> }
>
> s.host_start = target + bias;
Jakub
