1 files changed, 59 insertions, 16 deletions

diff --git a/kernel/sched/stats.h b/kernel/sched/stats.h
index 38f3698f5e5b..452826df6ae1 100644
--- a/kernel/sched/stats.h
+++ b/kernel/sched/stats.h
@@ -110,49 +110,84 @@ __schedstats_from_se(struct sched_entity *se)
 void psi_task_change(struct task_struct *task, int clear, int set);
 void psi_task_switch(struct task_struct *prev, struct task_struct *next,
 		     bool sleep);
-void psi_account_irqtime(struct task_struct *task, u32 delta);
-
+#ifdef CONFIG_IRQ_TIME_ACCOUNTING
+void psi_account_irqtime(struct rq *rq, struct task_struct *curr, struct task_struct *prev);
+#else
+static inline void psi_account_irqtime(struct rq *rq, struct task_struct *curr,
+				       struct task_struct *prev) {}
+#endif /*CONFIG_IRQ_TIME_ACCOUNTING */
 /*
  * PSI tracks state that persists across sleeps, such as iowaits and
  * memory stalls. As a result, it has to distinguish between sleeps,
- * where a task's runnable state changes, and requeues, where a task
- * and its state are being moved between CPUs and runqueues.
+ * where a task's runnable state changes, and migrations, where a task
+ * and its runnable state are being moved between CPUs and runqueues.
+ *
+ * A notable case is a task whose dequeue is delayed. PSI considers
+ * those sleeping, but because they are still on the runqueue they can
+ * go through migration requeues. In this case, *sleeping* states need
+ * to be transferred.
  */
-static inline void psi_enqueue(struct task_struct *p, bool wakeup)
+static inline void psi_enqueue(struct task_struct *p, int flags)
 {
-	int clear = 0, set = TSK_RUNNING;
+	int clear = 0, set = 0;
 
 	if (static_branch_likely(&psi_disabled))
 		return;
 
-	if (p->in_memstall)
-		set |= TSK_MEMSTALL_RUNNING;
+	/* Same runqueue, nothing changed for psi */
+	if (flags & ENQUEUE_RESTORE)
+		return;
+
+	/* psi_sched_switch() will handle the flags */
+	if (task_on_cpu(task_rq(p), p))
+		return;
 
-	if (!wakeup) {
+	if (p->se.sched_delayed) {
+		/* CPU migration of "sleeping" task */
+		WARN_ON_ONCE(!(flags & ENQUEUE_MIGRATED));
 		if (p->in_memstall)
 			set |= TSK_MEMSTALL;
+		if (p->in_iowait)
+			set |= TSK_IOWAIT;
+	} else if (flags & ENQUEUE_MIGRATED) {
+		/* CPU migration of runnable task */
+		set = TSK_RUNNING;
+		if (p->in_memstall)
+			set |= TSK_MEMSTALL | TSK_MEMSTALL_RUNNING;
 	} else {
+		/* Wakeup of new or sleeping task */
 		if (p->in_iowait)
 			clear |= TSK_IOWAIT;
+		set = TSK_RUNNING;
+		if (p->in_memstall)
+			set |= TSK_MEMSTALL_RUNNING;
 	}
 
 	psi_task_change(p, clear, set);
 }
 
-static inline void psi_dequeue(struct task_struct *p, bool sleep)
+static inline void psi_dequeue(struct task_struct *p, int flags)
 {
 	if (static_branch_likely(&psi_disabled))
 		return;
 
+	/* Same runqueue, nothing changed for psi */
+	if (flags & DEQUEUE_SAVE)
+		return;
+
 	/*
 	 * A voluntary sleep is a dequeue followed by a task switch. To
 	 * avoid walking all ancestors twice, psi_task_switch() handles
 	 * TSK_RUNNING and TSK_IOWAIT for us when it moves TSK_ONCPU.
 	 * Do nothing here.
 	 */
-	if (sleep)
+	if (flags & DEQUEUE_SLEEP)
 		return;
 
+	/*
+	 * When migrating a task to another CPU, clear all psi
+	 * state. The enqueue callback above will work it out.
+	 */
 	psi_task_change(p, p->psi_flags, 0);
 }
 
@@ -186,13 +221,14 @@ static inline void psi_sched_switch(struct task_struct *prev,
 }
 
 #else /* CONFIG_PSI */
-static inline void psi_enqueue(struct task_struct *p, bool wakeup) {}
-static inline void psi_dequeue(struct task_struct *p, bool sleep) {}
+static inline void psi_enqueue(struct task_struct *p, bool migrate) {}
+static inline void psi_dequeue(struct task_struct *p, bool migrate) {}
 static inline void psi_ttwu_dequeue(struct task_struct *p) {}
 static inline void psi_sched_switch(struct task_struct *prev,
 				    struct task_struct *next,
 				    bool sleep) {}
-static inline void psi_account_irqtime(struct task_struct *task, u32 delta) {}
+static inline void psi_account_irqtime(struct rq *rq, struct task_struct *curr,
+				       struct task_struct *prev) {}
 #endif /* CONFIG_PSI */
 
 #ifdef CONFIG_SCHED_INFO
@@ -212,14 +248,17 @@ static inline void sched_info_dequeue(struct rq *rq, struct task_struct *t)
 	delta = rq_clock(rq) - t->sched_info.last_queued;
 	t->sched_info.last_queued = 0;
 	t->sched_info.run_delay += delta;
-
+	if (delta > t->sched_info.max_run_delay)
+		t->sched_info.max_run_delay = delta;
+	if (delta && (!t->sched_info.min_run_delay || delta < t->sched_info.min_run_delay))
+		t->sched_info.min_run_delay = delta;
 	rq_sched_info_dequeue(rq, delta);
 }
 
 /*
  * Called when a task finally hits the CPU.  We can now calculate how
  * long it was waiting to run.  We also note when it began so that we
- * can keep stats on how long its timeslice is.
+ * can keep stats on how long its time-slice is.
  */
 static void sched_info_arrive(struct rq *rq, struct task_struct *t)
 {
@@ -234,6 +273,10 @@ static void sched_info_arrive(struct rq *rq, struct task_struct *t)
 	t->sched_info.run_delay += delta;
 	t->sched_info.last_arrival = now;
 	t->sched_info.pcount++;
+	if (delta > t->sched_info.max_run_delay)
+		t->sched_info.max_run_delay = delta;
+	if (delta && (!t->sched_info.min_run_delay || delta < t->sched_info.min_run_delay))
+		t->sched_info.min_run_delay = delta;
 
 	rq_sched_info_arrive(rq, delta);
 }