rcu: Add polled expedited grace-period primitives

This commit adds expedited grace-period functionality to RCU's polled
grace-period API, adding start_poll_synchronize_rcu_expedited() and
cond_synchronize_rcu_expedited(), which are similar to the existing
start_poll_synchronize_rcu() and cond_synchronize_rcu() functions,
respectively.

Note that although start_poll_synchronize_rcu_expedited() can be invoked
very early, the resulting expedited grace periods are not guaranteed
to start until after workqueues are fully initialized.  On the other
hand, both synchronize_rcu() and synchronize_rcu_expedited() can also
be invoked very early, and the resulting grace periods will be taken
into account as they occur.

[ paulmck: Apply feedback from Neeraj Upadhyay. ]

Link: https://lore.kernel.org/all/20220121142454.1994916-1-bfoster@redhat.com/
Link: https://docs.google.com/document/d/1RNKWW9jQyfjxw2E8dsXVTdvZYh0HnYeSHDKog9jhdN8/edit?usp=sharing
Cc: Brian Foster <bfoster@redhat.com>
Cc: Dave Chinner <david@fromorbit.com>
Cc: Al Viro <viro@zeniv.linux.org.uk>
Cc: Ian Kent <raven@themaw.net>
Signed-off-by: Paul E. McKenney <paulmck@kernel.org>

3 files changed, 104 insertions, 5 deletions

diff --git a/kernel/rcu/tree.c b/kernel/rcu/tree.c
index 1505b02b4e53..6cf5b51622cd 100644
--- a/kernel/rcu/tree.c
+++ b/kernel/rcu/tree.c
@@ -4027,20 +4027,20 @@ EXPORT_SYMBOL_GPL(poll_state_synchronize_rcu);
 /**
  * cond_synchronize_rcu - Conditionally wait for an RCU grace period
  *
- * @oldstate: value from get_state_synchronize_rcu() or start_poll_synchronize_rcu()
+ * @oldstate: value from get_state_synchronize_rcu(), start_poll_synchronize_rcu(), or start_poll_synchronize_rcu_expedited()
  *
  * If a full RCU grace period has elapsed since the earlier call to
  * get_state_synchronize_rcu() or start_poll_synchronize_rcu(), just return.
  * Otherwise, invoke synchronize_rcu() to wait for a full grace period.
  *
- * Yes, this function does not take counter wrap into account.  But
- * counter wrap is harmless.  If the counter wraps, we have waited for
+ * Yes, this function does not take counter wrap into account.
+ * But counter wrap is harmless.  If the counter wraps, we have waited for
  * more than 2 billion grace periods (and way more on a 64-bit system!),
- * so waiting for one additional grace period should be just fine.
+ * so waiting for a couple of additional grace periods should be just fine.
  *
  * This function provides the same memory-ordering guarantees that
  * would be provided by a synchronize_rcu() that was invoked at the call
- * to the function that provided @oldstate, and that returned at the end
+ * to the function that provided @oldstate and that returned at the end
  * of this function.
  */
 void cond_synchronize_rcu(unsigned long oldstate)
@@ -4793,6 +4793,9 @@ static void __init rcu_init_one(void)
 			init_waitqueue_head(&rnp->exp_wq[3]);
 			spin_lock_init(&rnp->exp_lock);
 			mutex_init(&rnp->boost_kthread_mutex);
+			raw_spin_lock_init(&rnp->exp_poll_lock);
+			rnp->exp_seq_poll_rq = RCU_GET_STATE_COMPLETED;
+			INIT_WORK(&rnp->exp_poll_wq, sync_rcu_do_polled_gp);
 		}
 	}
 
@@ -5018,6 +5021,10 @@ void __init rcu_init(void)
 		qovld_calc = DEFAULT_RCU_QOVLD_MULT * qhimark;
 	else
 		qovld_calc = qovld;
+
+	// Kick-start any polled grace periods that started early.
+	if (!(per_cpu_ptr(&rcu_data, cpu)->mynode->exp_seq_poll_rq & 0x1))
+		(void)start_poll_synchronize_rcu_expedited();
 }
 
 #include "tree_stall.h"
diff --git a/kernel/rcu/tree.h b/kernel/rcu/tree.h
index 5634e76106c4..fb77deca5f5c 100644
--- a/kernel/rcu/tree.h
+++ b/kernel/rcu/tree.h
@@ -133,6 +133,10 @@ struct rcu_node {
 	wait_queue_head_t exp_wq[4];
 	struct rcu_exp_work rew;
 	bool exp_need_flush;	/* Need to flush workitem? */
+	raw_spinlock_t exp_poll_lock;
+				/* Lock and data for polled expedited grace periods. */
+	unsigned long exp_seq_poll_rq;
+	struct work_struct exp_poll_wq;
 } ____cacheline_internodealigned_in_smp;
 
 /*
@@ -484,3 +488,6 @@ static void rcu_iw_handler(struct irq_work *iwp);
 static void check_cpu_stall(struct rcu_data *rdp);
 static void rcu_check_gp_start_stall(struct rcu_node *rnp, struct rcu_data *rdp,
 				     const unsigned long gpssdelay);
+
+/* Forward declarations for tree_exp.h. */
+static void sync_rcu_do_polled_gp(struct work_struct *wp);
diff --git a/kernel/rcu/tree_exp.h b/kernel/rcu/tree_exp.h
index e0258066b881..571b0a700cce 100644
--- a/kernel/rcu/tree_exp.h
+++ b/kernel/rcu/tree_exp.h
@@ -962,3 +962,88 @@ void synchronize_rcu_expedited(void)
 		synchronize_rcu_expedited_destroy_work(&rew);
 }
 EXPORT_SYMBOL_GPL(synchronize_rcu_expedited);
+
+/*
+ * Ensure that start_poll_synchronize_rcu_expedited() has the expedited
+ * RCU grace periods that it needs.
+ */
+static void sync_rcu_do_polled_gp(struct work_struct *wp)
+{
+	unsigned long flags;
+	struct rcu_node *rnp = container_of(wp, struct rcu_node, exp_poll_wq);
+	unsigned long s;
+
+	raw_spin_lock_irqsave(&rnp->exp_poll_lock, flags);
+	s = rnp->exp_seq_poll_rq;
+	rnp->exp_seq_poll_rq = RCU_GET_STATE_COMPLETED;
+	raw_spin_unlock_irqrestore(&rnp->exp_poll_lock, flags);
+	if (s == RCU_GET_STATE_COMPLETED)
+		return;
+	while (!poll_state_synchronize_rcu(s))
+		synchronize_rcu_expedited();
+	raw_spin_lock_irqsave(&rnp->exp_poll_lock, flags);
+	s = rnp->exp_seq_poll_rq;
+	if (poll_state_synchronize_rcu(s))
+		rnp->exp_seq_poll_rq = RCU_GET_STATE_COMPLETED;
+	raw_spin_unlock_irqrestore(&rnp->exp_poll_lock, flags);
+}
+
+/**
+ * start_poll_synchronize_rcu_expedited - Snapshot current RCU state and start expedited grace period
+ *
+ * Returns a cookie to pass to a call to cond_synchronize_rcu(),
+ * cond_synchronize_rcu_expedited(), or poll_state_synchronize_rcu(),
+ * allowing them to determine whether or not any sort of grace period has
+ * elapsed in the meantime.  If the needed expedited grace period is not
+ * already slated to start, initiates that grace period.
+ */
+unsigned long start_poll_synchronize_rcu_expedited(void)
+{
+	unsigned long flags;
+	struct rcu_data *rdp;
+	struct rcu_node *rnp;
+	unsigned long s;
+
+	s = get_state_synchronize_rcu();
+	rdp = per_cpu_ptr(&rcu_data, raw_smp_processor_id());
+	rnp = rdp->mynode;
+	if (rcu_init_invoked())
+		raw_spin_lock_irqsave(&rnp->exp_poll_lock, flags);
+	if (!poll_state_synchronize_rcu(s)) {
+		rnp->exp_seq_poll_rq = s;
+		if (rcu_init_invoked())
+			queue_work(rcu_gp_wq, &rnp->exp_poll_wq);
+	}
+	if (rcu_init_invoked())
+		raw_spin_unlock_irqrestore(&rnp->exp_poll_lock, flags);
+
+	return s;
+}
+EXPORT_SYMBOL_GPL(start_poll_synchronize_rcu_expedited);
+
+/**
+ * cond_synchronize_rcu_expedited - Conditionally wait for an expedited RCU grace period
+ *
+ * @oldstate: value from get_state_synchronize_rcu(), start_poll_synchronize_rcu(), or start_poll_synchronize_rcu_expedited()
+ *
+ * If any type of full RCU grace period has elapsed since the earlier
+ * call to get_state_synchronize_rcu(), start_poll_synchronize_rcu(),
+ * or start_poll_synchronize_rcu_expedited(), just return.  Otherwise,
+ * invoke synchronize_rcu_expedited() to wait for a full grace period.
+ *
+ * Yes, this function does not take counter wrap into account.
+ * But counter wrap is harmless.  If the counter wraps, we have waited for
+ * more than 2 billion grace periods (and way more on a 64-bit system!),
+ * so waiting for a couple of additional grace periods should be just fine.
+ *
+ * This function provides the same memory-ordering guarantees that
+ * would be provided by a synchronize_rcu() that was invoked at the call
+ * to the function that provided @oldstate and that returned at the end
+ * of this function.
+ */
+void cond_synchronize_rcu_expedited(unsigned long oldstate)
+{
+	if (!poll_state_synchronize_rcu(oldstate))
+		synchronize_rcu_expedited();
+}
+EXPORT_SYMBOL_GPL(cond_synchronize_rcu_expedited);