[RFC PATCH v4 5/9] sched, cpu hotplug: Use stable online cpus in try_to_wake_up() & select_task_rq()

["Srivatsa S. Bhat" <srivatsa.bhat@linux.vnet.ibm.com>]

Once stop_machine() is gone from the CPU offline path, we won't be able to
depend on preempt_disable() to prevent CPUs from going offline from under us.

Use the get/put_online_cpus_atomic() APIs to prevent CPUs from going offline,
while invoking from atomic context.

Scheduler functions such as try_to_wake_up() and select_task_rq() (and even
select_fallback_rq()) deal with picking new CPUs to run tasks. So they need
to synchronize with CPU offline operations.

Signed-off-by: Srivatsa S. Bhat <srivatsa.bhat@linux.vnet.ibm.com>
---

 kernel/sched/core.c |   14 ++++++++++++++
 1 file changed, 14 insertions(+)

diff --git a/kernel/sched/core.c b/kernel/sched/core.c
index 2d8927f..f51e0aa 100644
--- a/kernel/sched/core.c
+++ b/kernel/sched/core.c
@@ -1103,6 +1103,10 @@ EXPORT_SYMBOL_GPL(kick_process);
 #ifdef CONFIG_SMP
 /*
  * ->cpus_allowed is protected by both rq->lock and p->pi_lock
+ *
+ *  Must be called under get/put_online_cpus_atomic() or
+ *  equivalent, to avoid CPUs from going offline from underneath
+ *  us.
  */
 static int select_fallback_rq(int cpu, struct task_struct *p)
 {
@@ -1166,6 +1170,9 @@ out:
 
 /*
  * The caller (fork, wakeup) owns p->pi_lock, ->cpus_allowed is stable.
+ *
+ * Must be called under get/put_online_cpus_atomic(), to prevent
+ * CPUs from going offline from underneath us.
  */
 static inline
 int select_task_rq(struct task_struct *p, int sd_flags, int wake_flags)
@@ -1406,6 +1413,7 @@ try_to_wake_up(struct task_struct *p, unsigned int state, int wake_flags)
 	int cpu, success = 0;
 
 	smp_wmb();
+	get_online_cpus_atomic();
 	raw_spin_lock_irqsave(&p->pi_lock, flags);
 	if (!(p->state & state))
 		goto out;
@@ -1446,6 +1454,7 @@ stat:
 	ttwu_stat(p, cpu, wake_flags);
 out:
 	raw_spin_unlock_irqrestore(&p->pi_lock, flags);
+	put_online_cpus_atomic();
 
 	return success;
 }
@@ -1624,6 +1633,7 @@ void wake_up_new_task(struct task_struct *p)
 	unsigned long flags;
 	struct rq *rq;
 
+	get_online_cpus_atomic();
 	raw_spin_lock_irqsave(&p->pi_lock, flags);
 #ifdef CONFIG_SMP
 	/*
@@ -1644,6 +1654,7 @@ void wake_up_new_task(struct task_struct *p)
 		p->sched_class->task_woken(rq, p);
 #endif
 	task_rq_unlock(rq, p, &flags);
+	put_online_cpus_atomic();
 }
 
 #ifdef CONFIG_PREEMPT_NOTIFIERS
@@ -2541,6 +2552,7 @@ void sched_exec(void)
 	unsigned long flags;
 	int dest_cpu;
 
+	get_online_cpus_atomic();
 	raw_spin_lock_irqsave(&p->pi_lock, flags);
 	dest_cpu = p->sched_class->select_task_rq(p, SD_BALANCE_EXEC, 0);
 	if (dest_cpu == smp_processor_id())
@@ -2550,11 +2562,13 @@ void sched_exec(void)
 		struct migration_arg arg = { p, dest_cpu };
 
 		raw_spin_unlock_irqrestore(&p->pi_lock, flags);
+		put_online_cpus_atomic();
 		stop_one_cpu(task_cpu(p), migration_cpu_stop, &arg);
 		return;
 	}
 unlock:
 	raw_spin_unlock_irqrestore(&p->pi_lock, flags);
+	put_online_cpus_atomic();
 }
 
 #endif