On Fri, Oct 5, 2012 at 5:01 PM, Samuel Thibault <samuel.thibault@inria.fr> wrote:
Sebastian Kuzminsky, le Sat 06 Oct 2012 00:55:57 +0200, a écrit :
> binding to CPU0
> could not bind to CPU0: Resource deadlock avoided

Mmm, from what I read in the freebsd kernel:

/*
 * Create a set in the space provided in 'set' with the provided parameters.
 * The set is returned with a single ref.  May return EDEADLK if the set
 * will have no valid cpu based on restrictions from the parent.
 */

_cpuset_create(struct cpuset *set, struct cpuset *parent, const cpuset_t *mask,
    cpusetid_t id)
{

        if (!CPU_OVERLAP(&parent->cs_mask, mask))
                return (EDEADLK);

Could it be that due to administration rules lstopo is not allowed to
bind on cpu 0-9 ? In that case the x86 backend can not detect anything
there.


Hm.  It may be that we're doing something funny and reserving those CPUs.  I'll run some tests on Monday and get back to you.