rcu: Invoke rcu_spawn_core_kthreads() from rcu_spawn_gp_kthread()

[ Upstream commit 8e4b1d2 ]

Currently, rcu_spawn_core_kthreads() is invoked via an early_initcall(),
which works, except that rcu_spawn_gp_kthread() is also invoked via an
early_initcall() and rcu_spawn_core_kthreads() relies on adjustments to
kthread_prio that are carried out by rcu_spawn_gp_kthread().  There is
no guaranttee of ordering among early_initcall() handlers, and thus no
guarantee that kthread_prio will be properly checked and range-limited
at the time that rcu_spawn_core_kthreads() needs it.

In most cases, this bug is harmless.  After all, the only reason that
rcu_spawn_gp_kthread() adjusts the value of kthread_prio is if the user
specified a nonsensical value for this boot parameter, which experience
indicates is rare.

Nevertheless, a bug is a bug.  This commit therefore causes the
rcu_spawn_core_kthreads() function to be invoked directly from
rcu_spawn_gp_kthread() after any needed adjustments to kthread_prio have
been carried out.

Fixes: 48d07c0 ("rcu: Enable elimination of Tree-RCU softirq processing")
Signed-off-by: Paul E. McKenney <paulmck@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>

kernel/rcu/tree.c

@@ -2837,7 +2837,6 @@ static int __init rcu_spawn_core_kthreads(void)
 		  "%s: Could not start rcuc kthread, OOM is now expected behavior\n", __func__);
 	return 0;
 }
-early_initcall(rcu_spawn_core_kthreads);
 
 /*
  * Handle any core-RCU processing required by a call_rcu() invocation.
@@ -4273,6 +4272,7 @@ static int __init rcu_spawn_gp_kthread(void)
 	wake_up_process(t);
 	rcu_spawn_nocb_kthreads();
 	rcu_spawn_boost_kthreads();
+	rcu_spawn_core_kthreads();
 	return 0;
 }
 early_initcall(rcu_spawn_gp_kthread);