rcu: Add preemptibility checks in rcu_sched_qs() and rcu_bh_qs()
authorPaul E. McKenney <paulmck@linux.vnet.ibm.com>
Thu, 27 Apr 2017 23:13:53 +0000 (16:13 -0700)
committerPaul E. McKenney <paulmck@linux.vnet.ibm.com>
Thu, 8 Jun 2017 15:25:34 +0000 (08:25 -0700)
This commit adds WARN_ON_ONCE() calls that trigger if either
rcu_sched_qs() or rcu_bh_qs() are invoked with preemption enabled.
In the immortal words of Peter Zijlstra: "these are much harder to ignore
than comments".

Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
kernel/rcu/tree.c

index 3bee58fc23b169ec4413e4921caa622edca6a1d9..b01a02e7a0b7b3cbc11ae582c7d86c5e3b667566 100644 (file)
@@ -250,6 +250,7 @@ static int rcu_gp_in_progress(struct rcu_state *rsp)
  */
 void rcu_sched_qs(void)
 {
+       RCU_LOCKDEP_WARN(preemptible(), "rcu_sched_qs() invoked with preemption enabled!!!");
        if (!__this_cpu_read(rcu_sched_data.cpu_no_qs.s))
                return;
        trace_rcu_grace_period(TPS("rcu_sched"),
@@ -265,6 +266,7 @@ void rcu_sched_qs(void)
 
 void rcu_bh_qs(void)
 {
+       RCU_LOCKDEP_WARN(preemptible(), "rcu_bh_qs() invoked with preemption enabled!!!");
        if (__this_cpu_read(rcu_bh_data.cpu_no_qs.s)) {
                trace_rcu_grace_period(TPS("rcu_bh"),
                                       __this_cpu_read(rcu_bh_data.gpnum),