On Wed, 5 Apr 2017 21:15:15 -0700
"Paul E. McKenney" <paul...@linux.vnet.ibm.com> wrote:
\> > diff --git a/kernel/trace/ftrace.c b/kernel/trace/ftrace.c
> > index 8efd9fe..28e3019 100644
> > --- a/kernel/trace/ftrace.c
> > +++ b/kernel/trace/ftrace.c
> > @@ -2808,18 +2808,28 @@ static int ftrace_shutdown(struct ftrace_ops *ops,
> > int command)
> > * callers are done before leaving this function.
> > * The same goes for freeing the per_cpu data of the per_cpu
> > * ops.
> > - *
> > - * Again, normal synchronize_sched() is not good enough.
> > - * We need to do a hard force of sched synchronization.
> > - * This is because we use preempt_disable() to do RCU, but
> > - * the function tracers can be called where RCU is not watching
> > - * (like before user_exit()). We can not rely on the RCU
> > - * infrastructure to do the synchronization, thus we must do it
> > - * ourselves.
> > */
> > if (ops->flags & (FTRACE_OPS_FL_DYNAMIC | FTRACE_OPS_FL_PER_CPU)) {
> > + /*
> > + * We need to do a hard force of sched synchronization.
> > + * This is because we use preempt_disable() to do RCU, but
> > + * the function tracers can be called where RCU is not watching
> > + * (like before user_exit()). We can not rely on the RCU
> > + * infrastructure to do the synchronization, thus we must do it
> > + * ourselves.
> > + */
> > schedule_on_each_cpu(ftrace_sync);
>
> Great header comment on ftrace_sync(): "Yes, function tracing is rude."
> And schedule_on_each_cpu() looks like a great workqueue gatling gun! ;-)
>
> > +#ifdef CONFIG_PREEMPT
> > + /*
> > + * When the kernel is preeptive, tasks can be preempted
> > + * while on a ftrace trampoline. Just scheduling a task on
> > + * a CPU is not good enough to flush them. Calling
> > + * synchronize_rcu_tasks() will wait for those tasks to
> > + * execute and either schedule voluntarily or enter user space.
> > + */
> > + synchronize_rcu_tasks();
> > +#endif
>
> How about this to save a line?
>
> if (IS_ENABLED(CONFIG_PREEMPT))
> synchronize_rcu_tasks();
Ah, this works as gcc optimizes it out. Otherwise I received a compile
error with synchronize_rcu_tasks() not defined. But that's because I
never enabled CONFIG_TASKS_RCU.
>
> One thing that might speed this up a bit (or might not) would be to
> doe the schedule_on_each_cpu() from a delayed workqueue. That way,
> if any of the activity from schedule_on_each_cpu() involved a voluntary
> context switch (from a cond_resched() or some such), then
> synchronize_rcu_tasks() would get the benefit of that context switch.
>
> You would need a flush_work() to wait for that delayed workqueue
> as well, of course.
This is a very slow path, I'm not too interested in making it complex
to speed it up.
>
> Not sure whether it is worth it, but figured I should pass it along.
>
> > arch_ftrace_trampoline_free(ops);
> >
> > if (ops->flags & FTRACE_OPS_FL_PER_CPU)
> > @@ -5366,22 +5376,6 @@ void __weak arch_ftrace_update_trampoline(struct
> > ftrace_ops *ops)
> >
> > static void ftrace_update_trampoline(struct ftrace_ops *ops)
> > {
> > -
> > -/*
> > - * Currently there's no safe way to free a trampoline when the kernel
> > - * is configured with PREEMPT. That is because a task could be preempted
> > - * when it jumped to the trampoline, it may be preempted for a long time
> > - * depending on the system load, and currently there's no way to know
> > - * when it will be off the trampoline. If the trampoline is freed
> > - * too early, when the task runs again, it will be executing on freed
> > - * memory and crash.
> > - */
> > -#ifdef CONFIG_PREEMPT
> > - /* Currently, only non dynamic ops can have a trampoline */
> > - if (ops->flags & FTRACE_OPS_FL_DYNAMIC)
> > - return;
> > -#endif
> > -
> > arch_ftrace_update_trampoline(ops);
> > }
>
> Agreed, straightforward patch!
Great, I'll start making it official then.
-- Steve
