diff --git a/include/linux/rcupdate.h b/include/linux/rcupdate.h
index e6146d0074f8..f531b29207da 100644
--- a/include/linux/rcupdate.h
+++ b/include/linux/rcupdate.h
@@ -363,15 +363,20 @@ static inline void rcu_init_nohz(void)
 #ifdef CONFIG_TASKS_RCU
 #define TASKS_RCU(x) x
 extern struct srcu_struct tasks_rcu_exit_srcu;
-#define rcu_note_voluntary_context_switch(t) \
+#define rcu_note_voluntary_context_switch_lite(t) \
 	do { \
-		rcu_all_qs(); \
 		if (READ_ONCE((t)->rcu_tasks_holdout)) \
 			WRITE_ONCE((t)->rcu_tasks_holdout, false); \
 	} while (0)
+#define rcu_note_voluntary_context_switch(t) \
+	do { \
+		rcu_all_qs(); \
+		rcu_note_voluntary_context_switch_lite(t); \
+	} while (0)
 #else /* #ifdef CONFIG_TASKS_RCU */
 #define TASKS_RCU(x) do { } while (0)
-#define rcu_note_voluntary_context_switch(t)	rcu_all_qs()
+#define rcu_note_voluntary_context_switch_lite(t)	do { } while (0)
+#define rcu_note_voluntary_context_switch(t)		rcu_all_qs()
 #endif /* #else #ifdef CONFIG_TASKS_RCU */
 
 /**
diff --git a/include/linux/rcutiny.h b/include/linux/rcutiny.h
index 5219be250f00..74d9c3a1feee 100644
--- a/include/linux/rcutiny.h
+++ b/include/linux/rcutiny.h
@@ -92,10 +92,11 @@ static inline void kfree_call_rcu(struct rcu_head *head,
 	call_rcu(head, func);
 }
 
-static inline void rcu_note_context_switch(void)
-{
-	rcu_sched_qs();
-}
+#define rcu_note_context_switch(preempt) \
+	do { \
+		rcu_sched_qs(); \
+		rcu_note_voluntary_context_switch_lite(current); \
+	} while (0)
 
 /*
  * Take advantage of the fact that there is only one CPU, which
@@ -242,6 +243,10 @@ static inline bool rcu_is_watching(void)
 
 #endif /* #else defined(CONFIG_DEBUG_LOCK_ALLOC) || defined(CONFIG_RCU_TRACE) */
 
+static inline void rcu_request_urgent_qs_task(struct task_struct *t)
+{
+}
+
 static inline void rcu_all_qs(void)
 {
 	barrier(); /* Avoid RCU read-side critical sections leaking across. */
diff --git a/include/linux/rcutree.h b/include/linux/rcutree.h
index 63a4e4cf40a5..0bacb6b2af69 100644
--- a/include/linux/rcutree.h
+++ b/include/linux/rcutree.h
@@ -30,7 +30,7 @@
 #ifndef __LINUX_RCUTREE_H
 #define __LINUX_RCUTREE_H
 
-void rcu_note_context_switch(void);
+void rcu_note_context_switch(bool preempt);
 int rcu_needs_cpu(u64 basem, u64 *nextevt);
 void rcu_cpu_stall_reset(void);
 
@@ -41,7 +41,7 @@ void rcu_cpu_stall_reset(void);
  */
 static inline void rcu_virt_note_context_switch(int cpu)
 {
-	rcu_note_context_switch();
+	rcu_note_context_switch(false);
 }
 
 void synchronize_rcu_bh(void);
@@ -108,6 +108,7 @@ void rcu_scheduler_starting(void);
 extern int rcu_scheduler_active __read_mostly;
 
 bool rcu_is_watching(void);
+void rcu_request_urgent_qs_task(struct task_struct *t);
 
 void rcu_all_qs(void);
 
diff --git a/kernel/rcu/tree.c b/kernel/rcu/tree.c
index 3c23435d2083..891d97109e09 100644
--- a/kernel/rcu/tree.c
+++ b/kernel/rcu/tree.c
@@ -458,7 +458,7 @@ static void rcu_momentary_dyntick_idle(void)
  * and requires special handling for preemptible RCU.
  * The caller must have disabled interrupts.
  */
-void rcu_note_context_switch(void)
+void rcu_note_context_switch(bool preempt)
 {
 	barrier(); /* Avoid RCU read-side critical sections leaking down. */
 	trace_rcu_utilization(TPS("Start context switch"));
@@ -471,6 +471,8 @@ void rcu_note_context_switch(void)
 	if (unlikely(raw_cpu_read(rcu_dynticks.rcu_need_heavy_qs)))
 		rcu_momentary_dyntick_idle();
 	this_cpu_inc(rcu_dynticks.rcu_qs_ctr);
+	if (!preempt)
+		rcu_note_voluntary_context_switch_lite(current);
 out:
 	trace_rcu_utilization(TPS("End context switch"));
 	barrier(); /* Avoid RCU read-side critical sections leaking up. */
@@ -1149,6 +1151,24 @@ bool notrace rcu_is_watching(void)
 }
 EXPORT_SYMBOL_GPL(rcu_is_watching);
 
+/*
+ * If a holdout task is actually running, request an urgent quiescent
+ * state from its CPU.  This is unsynchronized, so migrations can cause
+ * the request to go to the wrong CPU.  Which is OK, all that will happen
+ * is that the CPU's next context switch will be a bit slower and next
+ * time around this task will generate another request.
+ */
+void rcu_request_urgent_qs_task(struct task_struct *t)
+{
+	int cpu;
+
+	barrier();
+	cpu = task_cpu(t);
+	if (!task_curr(t))
+		return; /* This task is not running on that CPU. */
+	smp_store_release(per_cpu_ptr(&rcu_dynticks.rcu_urgent_qs, cpu), true);
+}
+
 #if defined(CONFIG_PROVE_RCU) && defined(CONFIG_HOTPLUG_CPU)
 
 /*
diff --git a/kernel/rcu/update.c b/kernel/rcu/update.c
index c5df0d756900..273e869ca21d 100644
--- a/kernel/rcu/update.c
+++ b/kernel/rcu/update.c
@@ -665,6 +665,7 @@ static void check_holdout_task(struct task_struct *t,
 		put_task_struct(t);
 		return;
 	}
+	rcu_request_urgent_qs_task(t);
 	if (!needreport)
 		return;
 	if (*firstreport) {
diff --git a/kernel/sched/core.c b/kernel/sched/core.c
index 3b31fc05a0f1..2adf7b6c04e7 100644
--- a/kernel/sched/core.c
+++ b/kernel/sched/core.c
@@ -3378,7 +3378,7 @@ static void __sched notrace __schedule(bool preempt)
 		hrtick_clear(rq);
 
 	local_irq_disable();
-	rcu_note_context_switch();
+	rcu_note_context_switch(preempt);
 
 	/*
 	 * Make sure that signal_pending_state()->signal_pending() below