Re: [BUG] in sched fair on v3.13-rc1+ doing hotplug

[Peter Zijlstra]

On Wed, Nov 27, 2013 at 08:13:45AM -0500, Steven Rostedt wrote:
>         atomic_inc(&sd->groups->sgp->nr_busy_cpus);  <<---- BUG
> unlock:
>         rcu_read_unlock();
> }

tip/sched/urgent has this one:

---
commit 42eb088ed246a5a817bb45a8b32fe234cf1c0f8b
Author: Peter Zijlstra <peterz@xxxxxxxxxxxxx>
Date:   Tue Nov 19 16:41:49 2013 +0100

    sched: Avoid NULL dereference on sd_busy
    
    Commit 37dc6b50cee9 ("sched: Remove unnecessary iteration over sched
    domains to update nr_busy_cpus") forgot to clear 'sd_busy' under some
    conditions leading to a possible NULL deref in set_cpu_sd_state_idle().
    
    Reported-by: Anton Blanchard <anton@xxxxxxxxx>
    Cc: Preeti U Murthy <preeti@xxxxxxxxxxxxxxxxxx>
    Signed-off-by: Peter Zijlstra <peterz@xxxxxxxxxxxxx>
    Link: http://lkml.kernel.org/r/20131118113701.GF3866@xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx
    Signed-off-by: Ingo Molnar <mingo@xxxxxxxxxx>

diff --git a/kernel/sched/core.c b/kernel/sched/core.c
index c1808606ee5f..a1591ca7eb5a 100644
--- a/kernel/sched/core.c
+++ b/kernel/sched/core.c
@@ -4910,8 +4910,9 @@ static void update_top_cache_domain(int cpu)
 	if (sd) {
 		id = cpumask_first(sched_domain_span(sd));
 		size = cpumask_weight(sched_domain_span(sd));
-		rcu_assign_pointer(per_cpu(sd_busy, cpu), sd->parent);
+		sd = sd->parent; /* sd_busy */
 	}
+	rcu_assign_pointer(per_cpu(sd_busy, cpu), sd);
 
 	rcu_assign_pointer(per_cpu(sd_llc, cpu), sd);
 	per_cpu(sd_llc_size, cpu) = size;
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/