Open MPI logo

Hardware Locality Development Mailing List Archives

  |   Home   |   Support   |   FAQ   |   all Hardware Locality Development mailing list

Subject: Re: [hwloc-devel] [hwloc-svn] svn:hwloc r2649
From: Brice Goglin (Brice.Goglin_at_[hidden])
Date: 2010-10-28 04:13:48


This raises again a question that got no answer in previous emails.
Having NULL nodeset pointers in obj->*nodeset when the machine isn't
NUMA is annoying in many places. I think I will change to a full
nodeset instead. It will simplify many things.

Brice

Le 28/10/2010 10:09, bgoglin_at_[hidden] a écrit :
> Author: bgoglin
> Date: 2010-10-28 04:09:35 EDT (Thu, 28 Oct 2010)
> New Revision: 2649
> URL: https://svn.open-mpi.org/trac/hwloc/changeset/2649
>
> Log:
> Fix Linux get_thisthread_membind on non-numa machine
>
> hwloc_topology_get*_nodeset return NULL there, so return a full nodeset instead
>
>
> Text files modified:
> branches/membind/src/topology-linux.c | 6 +++++-
> 1 files changed, 5 insertions(+), 1 deletions(-)
>
> Modified: branches/membind/src/topology-linux.c
> ==============================================================================
> --- branches/membind/src/topology-linux.c (original)
> +++ branches/membind/src/topology-linux.c 2010-10-28 04:09:35 EDT (Thu, 28 Oct 2010)
> @@ -1057,7 +1057,11 @@
> goto out_with_mask;
>
> if (linuxpolicy == MPOL_DEFAULT) {
> - hwloc_bitmap_copy(nodeset, hwloc_topology_get_topology_nodeset(topology));
> + hwloc_const_nodeset_t topology_nodeset = hwloc_topology_get_topology_nodeset(topology);
> + if (topology_nodeset)
> + hwloc_bitmap_copy(nodeset, topology_nodeset);
> + else
> + hwloc_bitmap_fill(nodeset);
> } else {
> hwloc_linux_membind_mask_to_nodeset(topology, nodeset, max_os_index, linuxmask);
> }
> _______________________________________________
> hwloc-svn mailing list
> hwloc-svn_at_[hidden]
> http://www.open-mpi.org/mailman/listinfo.cgi/hwloc-svn
>