rcutorture: Adjust rcu_poll_need_2gp() for rcu_gp_oldstate field removal
authorPaul E. McKenney <paulmck@kernel.org>
Fri, 5 Aug 2022 01:05:46 +0000 (18:05 -0700)
committerPaul E. McKenney <paulmck@kernel.org>
Wed, 31 Aug 2022 12:09:21 +0000 (05:09 -0700)
Now that rcu_gp_oldstate can accurately track both normal and
expedited grace periods regardless of system state, rcutorture's
rcu_poll_need_2gp() function need only call for a second grace period
for the old single-unsigned-long grace-period polling APIs
This commit therefore adjusts rcu_poll_need_2gp() accordingly.

Signed-off-by: Paul E. McKenney <paulmck@kernel.org>
kernel/rcu/rcutorture.c

index 8995429..029de67 100644 (file)
@@ -520,7 +520,7 @@ static void rcu_sync_torture_init(void)
 
 static bool rcu_poll_need_2gp(bool poll, bool poll_full)
 {
-       return poll || (!IS_ENABLED(CONFIG_TINY_RCU) && poll_full && num_online_cpus() <= 1);
+       return poll;
 }
 
 static struct rcu_torture_ops rcu_ops = {