Move "force the use of hwloc" commits to v2.x

[jsquyres]

So far, these commits are involved:

• d97bc29
• f782a76
• 794ee4a

As of this writing (8am US Eastern Sep 5 2015), there's several PMIX-related conflicts when trying to merge the first of these commits to v2.x. I'm assuming the latest PMIX stuff hasn't been PR'ed over to v2.x -- when they are, these commits should apply cleanly.

[hppritcha]

@jsquyres do we care about the conflicts related to the lack of the treematch PRs in v2.x?

[hppritcha]

We probably need commit 7cb6c2f too.

[jsquyres]

7cb6c2f probably got swept up in open-mpi/ompi-release#671.

I'll file a PR for this one to start moving it forward.

[jsquyres]

@rhc54 I actually ran into a bunch of ORTE conflicts when I tried to merge the first commit into v2.x that are not immediately obvious to me how to fix... :-(

Can you have a look / take a shot at creating a PR?

        both modified:   opal/mca/btl/openib/btl_openib_component.c
        both modified:   orte/mca/ess/base/ess_base_std_orted.c
        both modified:   orte/mca/ess/hnp/ess_hnp_module.c
        both modified:   orte/mca/ess/pmi/ess_pmi_module.c
        both modified:   orte/mca/ras/simulator/ras_sim_module.c
        both modified:   orte/runtime/data_type_support/orte_dt_print_fns.c

[rhc54]

Okay - I'll give it a shot

[hppritcha]

It looks like we're missing commits, at least for orte/mca/ras/simulator/ras_sim_module.c:
ed93154
023936e

Also there were white-space changes that were not exactly equivalent.

[hppritcha]

I opened a PR open-mpi/ompi-release#729 to merge 023936e from 7/29 (or thereabouts) into ompi-release.

I would prefer to see 023936e and ed93154 go in as separate PRs to ompi-release.