rcu: Add READ_ONCE() to rcu_data ->gpwrap

The rcu_data structure's ->gpwrap field is read locklessly, and so
this commit adds the required READ_ONCE() to a pair of laods in order
to avoid destructive compiler optimizations.

This data race was reported by KCSAN.

Signed-off-by: Paul E. McKenney <paulmck@kernel.org>

kernel/rcu/tree.c

@@ -1322,7 +1322,7 @@ static void rcu_accelerate_cbs_unlocked(struct rcu_node *rnp,
 
 	rcu_lockdep_assert_cblist_protected(rdp);
 	c = rcu_seq_snap(&rcu_state.gp_seq);
-	if (!rdp->gpwrap && ULONG_CMP_GE(rdp->gp_seq_needed, c)) {
+	if (!READ_ONCE(rdp->gpwrap) && ULONG_CMP_GE(rdp->gp_seq_needed, c)) {
 		/* Old request still live, so mark recent callbacks. */
 		(void)rcu_segcblist_accelerate(&rdp->cblist, c);
 		return;

kernel/rcu/tree_stall.h

@@ -602,7 +602,7 @@ void show_rcu_gp_kthreads(void)
 			continue;
 		for_each_leaf_node_possible_cpu(rnp, cpu) {
 			rdp = per_cpu_ptr(&rcu_data, cpu);
-			if (rdp->gpwrap ||
+			if (READ_ONCE(rdp->gpwrap) ||
 			    ULONG_CMP_GE(READ_ONCE(rcu_state.gp_seq),
 					 READ_ONCE(rdp->gp_seq_needed)))
 				continue;