srcu: Move to state-based grace-period sequencing

The current SRCU grace-period processing might never reach the last
portion of srcu_advance_batches().  This is OK given the current
implementation, as the first portion, up to the try_check_zero()
following the srcu_flip() is sufficient to drive grace periods forward.
However, it has the unfortunate side-effect of making it impossible to
determine when a given grace period has ended, and it will be necessary
to efficiently trace ends of grace periods in order to efficiently handle
per-CPU SRCU callback lists.

This commit therefore adds states to the SRCU grace-period processing,
so that the end of a given SRCU grace period is marked by the transition
to the SRCU_STATE_DONE state.

Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>

2 files changed, 69 insertions, 52 deletions

diff --git a/include/linux/srcu.h b/include/linux/srcu.h
index a598cf3ac70c..f149a685896c 100644
--- a/include/linux/srcu.h
+++ b/include/linux/srcu.h
@@ -48,7 +48,7 @@ struct srcu_struct {
 	unsigned long completed;
 	struct srcu_array __percpu *per_cpu_ref;
 	spinlock_t queue_lock; /* protect ->batch_queue, ->running */
-	bool running;
+	int srcu_state;
 	/* callbacks just queued */
 	struct rcu_batch batch_queue;
 	/* callbacks try to do the first check_zero */
@@ -62,6 +62,12 @@ struct srcu_struct {
 #endif /* #ifdef CONFIG_DEBUG_LOCK_ALLOC */
 };
 
+/* Values for -> state variable. */
+#define SRCU_STATE_IDLE		0
+#define SRCU_STATE_SCAN1	1
+#define SRCU_STATE_SCAN2	2
+#define SRCU_STATE_DONE		3
+
 #ifdef CONFIG_DEBUG_LOCK_ALLOC
 
 int __init_srcu_struct(struct srcu_struct *sp, const char *name,
@@ -89,7 +95,7 @@ void process_srcu(struct work_struct *work);
 		.completed = -300,					\
 		.per_cpu_ref = &name##_srcu_array,			\
 		.queue_lock = __SPIN_LOCK_UNLOCKED(name.queue_lock),	\
-		.running = false,					\
+		.srcu_state = SRCU_STATE_IDLE,				\
 		.batch_queue = RCU_BATCH_INIT(name.batch_queue),	\
 		.batch_check0 = RCU_BATCH_INIT(name.batch_check0),	\
 		.batch_check1 = RCU_BATCH_INIT(name.batch_check1),	\
diff --git a/kernel/rcu/srcu.c b/kernel/rcu/srcu.c
index 821ecda873f2..70286f68f3a1 100644
--- a/kernel/rcu/srcu.c
+++ b/kernel/rcu/srcu.c
@@ -111,7 +111,7 @@ static int init_srcu_struct_fields(struct srcu_struct *sp)
 {
 	sp->completed = 0;
 	spin_lock_init(&sp->queue_lock);
-	sp->running = false;
+	sp->srcu_state = SRCU_STATE_IDLE;
 	rcu_batch_init(&sp->batch_queue);
 	rcu_batch_init(&sp->batch_check0);
 	rcu_batch_init(&sp->batch_check1);
@@ -270,7 +270,7 @@ void cleanup_srcu_struct(struct srcu_struct *sp)
 	if (WARN_ON(!rcu_all_batches_empty(sp)))
 		return; /* Leakage unless caller handles error. */
 	flush_delayed_work(&sp->work);
-	if (WARN_ON(sp->running))
+	if (WARN_ON(READ_ONCE(sp->srcu_state) != SRCU_STATE_IDLE))
 		return; /* Caller forgot to stop doing call_srcu()? */
 	free_percpu(sp->per_cpu_ref);
 	sp->per_cpu_ref = NULL;
@@ -391,8 +391,8 @@ void call_srcu(struct srcu_struct *sp, struct rcu_head *head,
 	spin_lock_irqsave(&sp->queue_lock, flags);
 	smp_mb__after_unlock_lock(); /* Caller's prior accesses before GP. */
 	rcu_batch_queue(&sp->batch_queue, head);
-	if (!sp->running) {
-		sp->running = true;
+	if (READ_ONCE(sp->srcu_state) == SRCU_STATE_IDLE) {
+		WRITE_ONCE(sp->srcu_state, SRCU_STATE_SCAN1);
 		queue_delayed_work(system_power_efficient_wq, &sp->work, 0);
 	}
 	spin_unlock_irqrestore(&sp->queue_lock, flags);
@@ -424,9 +424,9 @@ static void __synchronize_srcu(struct srcu_struct *sp, int trycount)
 	head->func = wakeme_after_rcu;
 	spin_lock_irq(&sp->queue_lock);
 	smp_mb__after_unlock_lock(); /* Caller's prior accesses before GP. */
-	if (!sp->running) {
+	if (READ_ONCE(sp->srcu_state) == SRCU_STATE_IDLE) {
 		/* steal the processing owner */
-		sp->running = true;
+		WRITE_ONCE(sp->srcu_state, SRCU_STATE_SCAN1);
 		rcu_batch_queue(&sp->batch_check0, head);
 		spin_unlock_irq(&sp->queue_lock);
 		/* give the processing owner to work_struct */
@@ -548,7 +548,9 @@ static void srcu_collect_new(struct srcu_struct *sp)
  */
 static void srcu_advance_batches(struct srcu_struct *sp, int trycount)
 {
-	int idx = 1 ^ (sp->completed & 1);
+	int idx;
+
+	WARN_ON_ONCE(sp->srcu_state == SRCU_STATE_IDLE);
 
 	/*
 	 * Because readers might be delayed for an extended period after
@@ -558,48 +560,56 @@ static void srcu_advance_batches(struct srcu_struct *sp, int trycount)
 	 * invoking a callback.
 	 */
 
-	if (rcu_batch_empty(&sp->batch_check0) &&
-	    rcu_batch_empty(&sp->batch_check1))
-		return; /* no callbacks need to be advanced */
-
-	if (!try_check_zero(sp, idx, trycount))
-		return; /* failed to advance, will try after SRCU_INTERVAL */
-
-	/*
-	 * The callbacks in ->batch_check1 have already done with their
-	 * first zero check and flip back when they were enqueued on
-	 * ->batch_check0 in a previous invocation of srcu_advance_batches().
-	 * (Presumably try_check_zero() returned false during that
-	 * invocation, leaving the callbacks stranded on ->batch_check1.)
-	 * They are therefore ready to invoke, so move them to ->batch_done.
-	 */
-	rcu_batch_move(&sp->batch_done, &sp->batch_check1);
-
-	if (rcu_batch_empty(&sp->batch_check0))
-		return; /* no callbacks need to be advanced */
-	srcu_flip(sp);
-
-	/*
-	 * The callbacks in ->batch_check0 just finished their
-	 * first check zero and flip, so move them to ->batch_check1
-	 * for future checking on the other idx.
-	 */
-	rcu_batch_move(&sp->batch_check1, &sp->batch_check0);
-
-	/*
-	 * SRCU read-side critical sections are normally short, so check
-	 * at least twice in quick succession after a flip.
-	 */
-	trycount = trycount < 2 ? 2 : trycount;
-	if (!try_check_zero(sp, idx^1, trycount))
-		return; /* failed to advance, will try after SRCU_INTERVAL */
+	if (sp->srcu_state == SRCU_STATE_DONE)
+		WRITE_ONCE(sp->srcu_state, SRCU_STATE_SCAN1);
+
+	if (sp->srcu_state == SRCU_STATE_SCAN1) {
+		idx = 1 ^ (sp->completed & 1);
+		if (!try_check_zero(sp, idx, trycount))
+			return; /* readers present, retry after SRCU_INTERVAL */
+
+		/*
+		 * The callbacks in ->batch_check1 have already done
+		 * with their first zero check and flip back when they were
+		 * enqueued on ->batch_check0 in a previous invocation of
+		 * srcu_advance_batches().  (Presumably try_check_zero()
+		 * returned false during that invocation, leaving the
+		 * callbacks stranded on ->batch_check1.) They are therefore
+		 * ready to invoke, so move them to ->batch_done.
+		 */
+		rcu_batch_move(&sp->batch_done, &sp->batch_check1);
+		srcu_flip(sp);
+
+		/*
+		 * The callbacks in ->batch_check0 just finished their
+		 * first check zero and flip, so move them to ->batch_check1
+		 * for future checking on the other idx.
+		 */
+		rcu_batch_move(&sp->batch_check1, &sp->batch_check0);
+
+		WRITE_ONCE(sp->srcu_state, SRCU_STATE_SCAN2);
+	}
 
-	/*
-	 * The callbacks in ->batch_check1 have now waited for all
-	 * pre-existing readers using both idx values.  They are therefore
-	 * ready to invoke, so move them to ->batch_done.
-	 */
-	rcu_batch_move(&sp->batch_done, &sp->batch_check1);
+	if (sp->srcu_state == SRCU_STATE_SCAN2) {
+
+		/*
+		 * SRCU read-side critical sections are normally short,
+		 * so check at least twice in quick succession after a flip.
+		 */
+		idx = 1 ^ (sp->completed & 1);
+		trycount = trycount < 2 ? 2 : trycount;
+		if (!try_check_zero(sp, idx, trycount))
+			return; /* readers present, retry after SRCU_INTERVAL */
+
+		/*
+		 * The callbacks in ->batch_check1 have now waited for
+		 * all pre-existing readers using both idx values.  They are
+		 * therefore ready to invoke, so move them to ->batch_done.
+		 */
+		rcu_batch_move(&sp->batch_done, &sp->batch_check1);
+
+		WRITE_ONCE(sp->srcu_state, SRCU_STATE_DONE);
+	}
 }
 
 /*
@@ -633,8 +643,9 @@ static void srcu_reschedule(struct srcu_struct *sp, unsigned long delay)
 
 	if (rcu_all_batches_empty(sp)) {
 		spin_lock_irq(&sp->queue_lock);
-		if (rcu_all_batches_empty(sp)) {
-			sp->running = false;
+		if (rcu_all_batches_empty(sp) &&
+		    READ_ONCE(sp->srcu_state) == SRCU_STATE_DONE) {
+			WRITE_ONCE(sp->srcu_state, SRCU_STATE_IDLE);
 			pending = false;
 		}
 		spin_unlock_irq(&sp->queue_lock);