sched/core: Fix DL runtime accounting during migration

Warnings are being triggered when running dl tasks from rt-app.
This seems to be a bug in walleye's 4.4 kernel arising from HMP
the revert. Due to that this patch was missed:

commit 3ea94de
Author: Joonwoo Park <joonwoop@codeaurora.org>
Date:   Thu Nov 12 19:38:54 2015 -0800
    sched/core: Fix incorrect wait time and wait count statistics

Fix the issue by reordering the flag setting before calling class
enqueue/dequeue callbacks.

Change-Id: I1b7ab408c307698ef67b0f1b05dcce06cc9ce625
Signed-off-by: Joel Fernandes <joelaf@google.com>

kernel/sched/core.c

@@ -1154,8 +1154,8 @@ static struct rq *move_queued_task(struct rq *rq, struct task_struct *p, int new
 {
 	lockdep_assert_held(&rq->lock);
 
-	dequeue_task(rq, p, 0);
 	p->on_rq = TASK_ON_RQ_MIGRATING;
+	dequeue_task(rq, p, 0);
 	double_lock_balance(rq, cpu_rq(new_cpu));
 	set_task_cpu(p, new_cpu);
 	double_unlock_balance(rq, cpu_rq(new_cpu));
@@ -1165,8 +1165,8 @@ static struct rq *move_queued_task(struct rq *rq, struct task_struct *p, int new
 
 	raw_spin_lock(&rq->lock);
 	BUG_ON(task_cpu(p) != new_cpu);
-	p->on_rq = TASK_ON_RQ_QUEUED;
 	enqueue_task(rq, p, 0);
+	p->on_rq = TASK_ON_RQ_QUEUED;
 	check_preempt_curr(rq, p, 0);
 
 	return rq;