Re: [RFC PATCH tip/master] kprobes: Use synchronize_rcu_tasks() for optprobe wit CONFIG_PREEMPT

[Ingo Molnar <mingo@kernel.org>]

* Masami Hiramatsu <mhiramat@kernel.org> wrote:

> --- a/kernel/kprobes.c
> +++ b/kernel/kprobes.c
> @@ -377,6 +377,23 @@ static inline void copy_kprobe(struct kprobe *ap, struct kprobe *p)
>  static bool kprobes_allow_optimization;
>  
>  /*
> + * Synchronizing wait on trampline code for interrupted tasks/threads.
> + * Since the threads running on dynamically allocated trampline code
> + * can be interrupted, kprobes has to wait for those tasks back on
> + * track and scheduled. If the kernel is preemptive, the thread can be
> + * preempted by other tasks on the trampoline too. For such case, this
> + * calls synchronize_rcu_tasks() to wait for those tasks back on track.
> + */
> +static void synchronize_on_trampoline(void)
> +{
> +#ifdef CONFIG_PREEMPT
> +	synchronize_rcu_tasks();
> +#else
> +	synchronize_sched();
> +#endif
> +}

So that's really unacceptably ugly.

Paul, I still question the need to have tasks-RCU as a Kconfig distinction, 
_especially_ if its API usage results in such ugly secondary #ifdefs...

Why isn't there a single synchronize_rcu_tasks() API function, which does what is 
expected, where the _RCU_ code figures out how to implement it?

I.e.:

 - There should be no user configurable TASKS_RCU Kconfig setting - at most a
   helper Kconfig that is automatically selected by the RCU code itself.

 - Both ftrace andkprobes would use a simple synchronize_rcu_tasks() call.

Thanks,

	Ingo