rcu: Clarify memory-ordering properties of grace-period primitives

This commit explicitly states the memory-ordering properties of the
RCU grace-period primitives.  Although these properties were in some
sense implied by the fundmental property of RCU ("a grace period must
wait for all pre-existing RCU read-side critical sections to complete"),
stating it explicitly will be a great labor-saving device.

Reported-by: Oleg Nesterov <oleg@redhat.com>
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Reviewed-by: Oleg Nesterov <oleg@redhat.com>

3 files changed, 58 insertions, 4 deletions

diff --git a/include/linux/rcupdate.h b/include/linux/rcupdate.h
index 7c968e4f..6256759 100644
--- a/include/linux/rcupdate.h
+++ b/include/linux/rcupdate.h
@@ -90,6 +90,25 @@ extern void do_trace_rcu_torture_read(char *rcutorturename,
  * that started after call_rcu() was invoked.  RCU read-side critical
  * sections are delimited by rcu_read_lock() and rcu_read_unlock(),
  * and may be nested.
+ *
+ * Note that all CPUs must agree that the grace period extended beyond
+ * all pre-existing RCU read-side critical section.  On systems with more
+ * than one CPU, this means that when "func()" is invoked, each CPU is
+ * guaranteed to have executed a full memory barrier since the end of its
+ * last RCU read-side critical section whose beginning preceded the call
+ * to call_rcu().  It also means that each CPU executing an RCU read-side
+ * critical section that continues beyond the start of "func()" must have
+ * executed a memory barrier after the call_rcu() but before the beginning
+ * of that RCU read-side critical section.  Note that these guarantees
+ * include CPUs that are offline, idle, or executing in user mode, as
+ * well as CPUs that are executing in the kernel.
+ *
+ * Furthermore, if CPU A invoked call_rcu() and CPU B invoked the
+ * resulting RCU callback function "func()", then both CPU A and CPU B are
+ * guaranteed to execute a full memory barrier during the time interval
+ * between the call to call_rcu() and the invocation of "func()" -- even
+ * if CPU A and CPU B are the same CPU (but again only if the system has
+ * more than one CPU).
  */
 extern void call_rcu(struct rcu_head *head,
 			      void (*func)(struct rcu_head *head));
@@ -118,6 +137,9 @@ extern void call_rcu(struct rcu_head *head,
  *  OR
  *  - rcu_read_lock_bh() and rcu_read_unlock_bh(), if in process context.
  *  These may be nested.
+ *
+ * See the description of call_rcu() for more detailed information on
+ * memory ordering guarantees.
  */
 extern void call_rcu_bh(struct rcu_head *head,
 			void (*func)(struct rcu_head *head));
@@ -137,6 +159,9 @@ extern void call_rcu_bh(struct rcu_head *head,
  *  OR
  *  anything that disables preemption.
  *  These may be nested.
+ *
+ * See the description of call_rcu() for more detailed information on
+ * memory ordering guarantees.
  */
 extern void call_rcu_sched(struct rcu_head *head,
 			   void (*func)(struct rcu_head *rcu));
diff --git a/kernel/rcutree.c b/kernel/rcutree.c
index e4c2192..15a2bee 100644
--- a/kernel/rcutree.c
+++ b/kernel/rcutree.c
@@ -2228,10 +2228,28 @@ static inline int rcu_blocking_is_gp(void)
  * rcu_read_lock_sched().
  *
  * This means that all preempt_disable code sequences, including NMI and
- * hardware-interrupt handlers, in progress on entry will have completed
- * before this primitive returns.  However, this does not guarantee that
- * softirq handlers will have completed, since in some kernels, these
- * handlers can run in process context, and can block.
+ * non-threaded hardware-interrupt handlers, in progress on entry will
+ * have completed before this primitive returns.  However, this does not
+ * guarantee that softirq handlers will have completed, since in some
+ * kernels, these handlers can run in process context, and can block.
+ *
+ * Note that this guarantee implies further memory-ordering guarantees.
+ * On systems with more than one CPU, when synchronize_sched() returns,
+ * each CPU is guaranteed to have executed a full memory barrier since the
+ * end of its last RCU-sched read-side critical section whose beginning
+ * preceded the call to synchronize_sched().  In addition, each CPU having
+ * an RCU read-side critical section that extends beyond the return from
+ * synchronize_sched() is guaranteed to have executed a full memory barrier
+ * after the beginning of synchronize_sched() and before the beginning of
+ * that RCU read-side critical section.  Note that these guarantees include
+ * CPUs that are offline, idle, or executing in user mode, as well as CPUs
+ * that are executing in the kernel.
+ *
+ * Furthermore, if CPU A invoked synchronize_sched(), which returned
+ * to its caller on CPU B, then both CPU A and CPU B are guaranteed
+ * to have executed a full memory barrier during the execution of
+ * synchronize_sched() -- even if CPU A and CPU B are the same CPU (but
+ * again only if the system has more than one CPU).
  *
  * This primitive provides the guarantees made by the (now removed)
  * synchronize_kernel() API.  In contrast, synchronize_rcu() only
@@ -2259,6 +2277,9 @@ EXPORT_SYMBOL_GPL(synchronize_sched);
  * read-side critical sections have completed.  RCU read-side critical
  * sections are delimited by rcu_read_lock_bh() and rcu_read_unlock_bh(),
  * and may be nested.
+ *
+ * See the description of synchronize_sched() for more detailed information
+ * on memory ordering guarantees.
  */
 void synchronize_rcu_bh(void)
 {
diff --git a/kernel/rcutree_plugin.h b/kernel/rcutree_plugin.h
index f921154..57e0ef8 100644
--- a/kernel/rcutree_plugin.h
+++ b/kernel/rcutree_plugin.h
@@ -670,6 +670,9 @@ EXPORT_SYMBOL_GPL(kfree_call_rcu);
  * concurrently with new RCU read-side critical sections that began while
  * synchronize_rcu() was waiting.  RCU read-side critical sections are
  * delimited by rcu_read_lock() and rcu_read_unlock(), and may be nested.
+ *
+ * See the description of synchronize_sched() for more detailed information
+ * on memory ordering guarantees.
  */
 void synchronize_rcu(void)
 {
@@ -875,6 +878,11 @@ EXPORT_SYMBOL_GPL(synchronize_rcu_expedited);
 
 /**
  * rcu_barrier - Wait until all in-flight call_rcu() callbacks complete.
+ *
+ * Note that this primitive does not necessarily wait for an RCU grace period
+ * to complete.  For example, if there are no RCU callbacks queued anywhere
+ * in the system, then rcu_barrier() is within its rights to return
+ * immediately, without waiting for anything, much less an RCU grace period.
  */
 void rcu_barrier(void)
 {