[PATCH rcu 06/12] rcu: Make Tiny RCU grace periods visible to polled APIs

[Paul E. McKenney]

This commit makes the Tiny RCU implementation of synchronize_rcu()
increment the rcu_ctrlblk.gp_seq counter, thus making both
synchronize_rcu() and synchronize_rcu_expedited() visible to
get_state_synchronize_rcu() and friends.

Link: https://lore.kernel.org/all/20220121142454.1994916-1-bfoster@xxxxxxxxxx/
Link: https://docs.google.com/document/d/1RNKWW9jQyfjxw2E8dsXVTdvZYh0HnYeSHDKog9jhdN8/edit?usp=sharing
Cc: Brian Foster <bfoster@xxxxxxxxxx>
Cc: Dave Chinner <david@xxxxxxxxxxxxx>
Cc: Al Viro <viro@xxxxxxxxxxxxxxxxxx>
Cc: Ian Kent <raven@xxxxxxxxxx>
Signed-off-by: Paul E. McKenney <paulmck@xxxxxxxxxx>
---
 kernel/rcu/tiny.c | 7 +++++--
 1 file changed, 5 insertions(+), 2 deletions(-)

diff --git a/kernel/rcu/tiny.c b/kernel/rcu/tiny.c
index dbee6bea67269..60071817d9399 100644
--- a/kernel/rcu/tiny.c
+++ b/kernel/rcu/tiny.c
@@ -139,8 +139,10 @@ static __latent_entropy void rcu_process_callbacks(struct softirq_action *unused
 /*
  * Wait for a grace period to elapse.  But it is illegal to invoke
  * synchronize_rcu() from within an RCU read-side critical section.
- * Therefore, any legal call to synchronize_rcu() is a quiescent
- * state, and so on a UP system, synchronize_rcu() need do nothing.
+ * Therefore, any legal call to synchronize_rcu() is a quiescent state,
+ * and so on a UP system, synchronize_rcu() need do nothing, other than
+ * let the polled APIs know that another grace period elapsed.
+ *
  * (But Lai Jiangshan points out the benefits of doing might_sleep()
  * to reduce latency.)
  *
@@ -152,6 +154,7 @@ void synchronize_rcu(void)
 			 lock_is_held(&rcu_lock_map) ||
 			 lock_is_held(&rcu_sched_lock_map),
 			 "Illegal synchronize_rcu() in RCU read-side critical section");
+	WRITE_ONCE(rcu_ctrlblk.gp_seq, rcu_ctrlblk.gp_seq + 2);
 }
 EXPORT_SYMBOL_GPL(synchronize_rcu);
 
-- 
2.31.1.189.g2e36527f23