Re: WARNING: cpu_is_offline() at native_smp_send_reschedule()

[Peter Zijlstra <a.p.zijlstra@chello.nl>]

On Fri, 2012-09-07 at 09:20 +0800, Fengguang Wu wrote:

> FYI, the bisect result is
> 
> commit 554cecaf733623b327eef9652b65965eb1081b81
> Author: Diwakar Tundlam <dtundlam@nvidia.com>
> Date:   Wed Mar 7 14:44:26 2012 -0800
> 
>     sched/nohz: Correctly initialize 'next_balance' in 'nohz' idle balancer
> 
>     The 'next_balance' field of 'nohz' idle balancer must be initialized
>     to jiffies. Since jiffies is initialized to negative 300 seconds the
>     'nohz' idle balancer does not run for the first 300s (5mins) after
>     bootup. If no new processes are spawed or no idle cycles happen, the
>     load on the cpus will remain unbalanced for that duration.
> 
>     Signed-off-by: Diwakar Tundlam <dtundlam@nvidia.com>
>     Signed-off-by: Peter Zijlstra <a.p.zijlstra@chello.nl>
>     Link: http://lkml.kernel.org/r/1DD7BFEDD3147247B1355BEFEFE4665237994F30EF@HQMAIL04.nvidia.com
>     Signed-off-by: Ingo Molnar <mingo@elte.hu>

Oh fun.. does the below 'fix' it?

The thing I'm thinking of a tick happening right after we set jiffies
but before the zalloc (specifically the memset(0)) is complete. Since
we've already registered the softirq we can end up in the load-balancer
and see a completely weird idle mask.

Hmm?

---
 kernel/sched/fair.c | 5 ++---
 1 file changed, 2 insertions(+), 3 deletions(-)

diff --git a/kernel/sched/fair.c b/kernel/sched/fair.c
index 1ca4fe4..ac57bb6 100644
--- a/kernel/sched/fair.c
+++ b/kernel/sched/fair.c
@@ -5346,13 +5346,12 @@ void print_cfs_stats(struct seq_file *m, int
cpu)
 __init void init_sched_fair_class(void)
 {
 #ifdef CONFIG_SMP
-	open_softirq(SCHED_SOFTIRQ, run_rebalance_domains);
-
 #ifdef CONFIG_NO_HZ
 	nohz.next_balance = jiffies;
 	zalloc_cpumask_var(&nohz.idle_cpus_mask, GFP_NOWAIT);
 	cpu_notifier(sched_ilb_notifier, 0);
 #endif
-#endif /* SMP */
 
+	open_softirq(SCHED_SOFTIRQ, run_rebalance_domains);
+#endif /* SMP */
 }