On Mon 2022-06-27 23:56 -0400, Steven Rostedt wrote: > > + preempt_enable(); > > ^^^^^^^^^^^^^^^^^ > > This just introduced an unbounded latency and will fail PREEMPT_RT. > Hi Steve, Sorry about that! If I understand correctly, there is absolutely no need to explicitly disable preemption in this context. I will use: rcu_read_lock() and rcu_read_unlock(), respectively, which will then be suitable for the preemptible RCU implementation. Kind regards, -- Aaron Tomlin