Re: [PATCH tip/core/rcu 3/4] rcu-tasks: Make ksoftirqd provide RCU Tasks quiescent states

[Paul E. McKenney]

On Fri, May 14, 2021 at 03:04:31PM +0900, Masami Hiramatsu wrote:
> Hi Paul,
> 
> On Thu, 13 May 2021 12:15:39 -0700
> "Paul E. McKenney" <paulmck@xxxxxxxxxx> wrote:
> 
> > On Fri, May 14, 2021 at 02:49:12AM +0900, Masami Hiramatsu wrote:
> > > On Thu, 13 May 2021 07:21:10 -0700
> > > "Paul E. McKenney" <paulmck@xxxxxxxxxx> wrote:
> > > 
> > > > On Thu, May 13, 2021 at 03:54:17PM +0900, Masami Hiramatsu wrote:
> > > > > Hi Paul,
> > > > > 
> > > > > On Wed, 12 May 2021 11:27:46 -0700
> > > > > "Paul E. McKenney" <paulmck@xxxxxxxxxx> wrote:
> > > > > 
> > > > > > Heavy networking load can cause a CPU to execute continuously and
> > > > > > indefinitely within ksoftirqd, in which case there will be no voluntary
> > > > > > task switches and thus no RCU-tasks quiescent states.  This commit
> > > > > > therefore causes the exiting rcu_softirq_qs() to provide an RCU-tasks
> > > > > > quiescent state.
> > > > > > 
> > > > > > This of course means that __do_softirq() and its callers cannot be
> > > > > > invoked from within a tracing trampoline.
> > > > > 
> > > > > I would like to confirm that you mean "tracing trampoline" here is
> > > > > the code on the trampoline buffer, not the handler code which is
> > > > > invoked from the trampoline buffer but it is protected by preempt_disable(),
> > > > > am I understand correctly?
> > > > 
> > > > Maybe?  ;-)
> > > > 
> > > > If the handler code is invoked from the trampoline buffer, but
> > > > returns somewhere else, then it is OK for the handler code to invoke
> > > > __do_softirq() or its callers.
> > > > 
> > > > In addition, if the handler code is invoked from the trampoline buffer is
> > > > guaranteed never to be running in the context of the ksoftirqd kthread,
> > > > then it is also OK for the handler code to invoke __do_softirq() or
> > > > its callers.
> > > > 
> > > > Otherwise, if the handler code might return back into the trampoline
> > > > buffer and if that code might be running in the context of the ksoftirqd
> > > > kthread, invoking __do_softirq() or one of its callers could result in
> > > > the trampoline buffer no longer being there when it was returned to.
> > > 
> > > Hmm, the optprobe may be involved in this case. It always return to
> > > the trampoline and handler does not disable irqs (only disable preempt).
> > > BTW, what will call the __do_softirq()? Is hardirq safe?
> > 
> > As long as your code does not explicitly call __do_softirq() or one of
> > its callers, you should be OK.
> > 
> > Let's suppose that your code takes a hardirq from ksoftirqd context.
> > In that case, the return-from-irq path will notice the ksoftirqd
> > context and refrain from calling __do_softirqd().  Life is good.
> > (See the invoke_softirq() function for more detail.)
> > 
> > On the other hand, if your code takes a hardirq from some non-ksoftirqd
> > context, and if this hardirq decides to handle softirqs on exit
> > from the hardirq, the "__this_cpu_read(ksoftirqd) == current" within
> > __do_softirq() will fail, so that rcu_softirq_qs() will not be called.
> > Life is still good.
> 
> Ah, OK. This is good.
> 
> > 
> > Either way, as long as your handler does not explicitly invoke
> > __do_softirq(), life is good.
> 
> There should be no such code, I hope. 
> 
> > 
> > The bad case is when you instrument a function that is invoked in the
> > context of a ksoftirqd kthread, and the corresponding handler (or
> > some function that the handler explicitly calls) directly invokes
> > __do_softirq() or one of its caller.
> > 
> > Is that more helpful?
> 
> OK, I got it. So it would be better to be commented later.
> But anyway I can't imagine that there is any reason to call
> __do_softirq() inside kprobe handler :)
> 
> Reviewed-by: Masami Hiramatsu <mhiramat@xxxxxxxxxx>

I will apply on the next rebase, thank you!

							Thanx, Paul