sched: Always initialize cpu-power

Often when we run into mis-shapen topologies the balance iteration
fails to update the cpu power properly and we'll end up in /0 traps.

Always initialize the cpu-power to a semi-sane value so that we can
at least boot the machine, even if the load-balancer might not
function correctly.

Signed-off-by: Peter Zijlstra <a.p.zijlstra@chello.nl>
Link: http://lkml.kernel.org/n/tip-3lbhyj25sr169ha7z3qht5na@git.kernel.org
Signed-off-by: Ingo Molnar <mingo@kernel.org>

diff --git a/kernel/sched/core.c b/kernel/sched/core.c
index 781acb9..725ee7c 100644 (file)
--- a/kernel/sched/core.c
+++ b/kernel/sched/core.c
@@ -5604,7 +5604,12 @@ static int sched_domain_debug_one(struct sched_domain *sd, int cpu, int level,
                        break;
                }
 
-               if (!group->sgp->power) {
+               /*
+                * Even though we initialize ->power to something semi-sane,
+                * we leave power_orig unset. This allows us to detect if
+                * domain iteration is still funny without causing /0 traps.
+                */
+               if (!group->sgp->power_orig) {
                        printk(KERN_CONT "\n");
                        printk(KERN_ERR "ERROR: domain->cpu_power not "
                                        "set\n");
@@ -6075,6 +6080,12 @@ build_overlap_sched_groups(struct sched_domain *sd, int cpu)
                if (atomic_inc_return(&sg->sgp->ref) == 1)
                        build_group_mask(sd, sg);
 
+               /*
+                * Initialize sgp->power such that even if we mess up the
+                * domains and no possible iteration will get us here, we won't
+                * die on a /0 trap.
+                */
+               sg->sgp->power = SCHED_POWER_SCALE * cpumask_weight(sg_span);
 
                /*
                 * Make sure the first group of this domain contains the

diff --git a/kernel/sched/fair.c b/kernel/sched/fair.c
index 54cbaa4..c9fd6d6 100644 (file)
--- a/kernel/sched/fair.c
+++ b/kernel/sched/fair.c
@@ -3602,7 +3602,7 @@ void update_group_power(struct sched_domain *sd, int cpu)
                } while (group != child->groups);
        }
 
-       sdg->sgp->power = power;
+       sdg->sgp->power_orig = sdg->sgp->power = power;
 }
 
 /*