[gcc/devel/omp/gcc-16] OpenMP: Add omp_get_device_distances routine: Initialize NUMA node for NVIDIA GPUs [PR125877]

https://gcc.gnu.org/g:a205e27a8bdca56771926520d8840fbd451713ff

commit a205e27a8bdca56771926520d8840fbd451713ff
Author: Thomas Schwinge <[email protected]>
Date:   Tue Jun 30 08:55:08 2026 +0200

    OpenMP: Add omp_get_device_distances routine: Initialize NUMA node for NVIDIA GPUs [PR125877]
    
    Fix-up for commit 8af1592882509fbead16e30fd2d056330a9609e4
    "OpenMP: Add omp_get_device_distances routine [PR125877]".  Without this,
    'libgomp/plugin/plugin-nvptx.c:GOMP_OFFLOAD_get_numa_node' returns random junk:
    
        /* Initialized to 0; to distinguish, save with offset.  */
        if (dev->numa_node != 0)
          return dev->numa_node > 0 ? dev->numa_node - 1 : dev->numa_node;
    
            PR libgomp/125877
            libgomp/
            * plugin/plugin-nvptx.c (nvptx_open_device): Initialize 'numa_node'.
    
    (cherry picked from commit 7759d1aaa500e36de9f87c0dc5e453787e07f051)

Diff:
---
 libgomp/plugin/plugin-nvptx.c | 2 ++
 1 file changed, 2 insertions(+)

diff --git a/libgomp/plugin/plugin-nvptx.c b/libgomp/plugin/plugin-nvptx.c
index e6a5d880b171..2f644a0acfcf 100644
--- a/libgomp/plugin/plugin-nvptx.c
+++ b/libgomp/plugin/plugin-nvptx.c
@@ -561,6 +561,8 @@ nvptx_open_device (int n)
 			 CU_DEVICE_ATTRIBUTE_UNIFIED_ADDRESSING, dev);
   assert (r == CUDA_SUCCESS && pi);
 
+  ptx_dev->numa_node = 0;
+
   for (int i = 0; i != GOMP_DIM_MAX; i++)
     ptx_dev->default_dims[i] = 0;