Re: [PATCH v12 3/3] tracing: Centralize preemptirq tracepoints and unify their usage

From: Paul E. McKenney
Date: Wed Aug 08 2018 - 10:49:32 EST


On Wed, Aug 08, 2018 at 07:10:53AM -0700, Joel Fernandes wrote:
> On Wed, Aug 8, 2018 at 6:00 AM, Paul E. McKenney
> <paulmck@xxxxxxxxxxxxxxxxxx> wrote:
> > On Tue, Aug 07, 2018 at 08:53:54PM -0700, Joel Fernandes wrote:
> >> On Tue, Aug 7, 2018 at 8:44 PM, Joel Fernandes <joelaf@xxxxxxxxxx> wrote:
> >> > Hi Steve,
> >> >
> >> > On Tue, Aug 7, 2018 at 7:28 PM, Steven Rostedt <rostedt@xxxxxxxxxxx> wrote:
> >> [...]
> >> >>> @@ -171,8 +174,7 @@ extern void syscall_unregfunc(void);
> >> >>> } while ((++it_func_ptr)->func); \
> >> >>> } \
> >> >>> \
> >> >>> - if (rcuidle) \
> >> >>> - srcu_read_unlock_notrace(&tracepoint_srcu, idx);\
> >> >>> + srcu_read_unlock_notrace(ss, idx); \
> >> >>
> >> >> Hmm, why do we have the two different srcu handles?
> >> >
> >> > Because if the memory operations happening on the normal SRCU handle
> >> > (during srcu_read_lock) is interrupted by NMI, then the other handle
> >> > (devoted to NMI) could be used instead and not bother the interrupted
> >> > handle. Does that makes sense?
> >> >
> >> > When I talked to Paul few months ago about SRCU from NMI context, he
> >> > mentioned the per-cpu memory operations during srcu_read_lock can be
> >> > NMI interrupted, that's why we added that warning.
> >>
> >> So I looked more closely, __srcu_read_lock on 2 different handles may
> >> still be doing a this_cpu_inc on the same location..
> >> (sp->sda->srcu_lock_count). :-(
> >>
> >> Paul any ideas on how to solve this?
> >
> > You lost me on this one. When you said "2 different handles", I assumed
> > that you meant two different values of "sp", which would have two
> > different addresses for &sp->sda->srcu_lock_count. What am I missing?
>
> Thanks a lot for the reply.
> I thought "sda" is the same for different srcu_struct(s). May be it
> was too late for me in the night, that's why I thought so? Which makes
> no sense now that I think of it.

I know that feeling! ;-)

> In that case based on what you're saying, the patch I sent to using
> different srcu_struct for NMI is still good I guess...

As long as you wait for both SRCU grace periods. Hmmm... Maybe that means
that there is still a use for synchronize_rcu_mult():

void call_srcu_nmi(struct rcu_head *rhp, rcu_callback_t func)
{
call_srcu(&trace_srcu_struct_nmi, rhp, func);
}

void call_srcu_nonmi(struct rcu_head *rhp, rcu_callback_t func)
{
call_srcu(&trace_srcu_struct_nonmi, rhp, func);
}

...

/* Wait concurrently on the two grace periods. */
synchronize_rcu_mult(call_srcu_nmi, call_srcu_nonmi);

On the other hand, I bet that doing this is just fine in your use case:

synchronize_srcu(&trace_srcu_struct_nmi);
synchronize_srcu(&trace_srcu_struct_nonmi);

But please note that synchronize_rcu_mult() is no more in my -rcu tree,
so if you do want it please let me know (and please let me know why it
is important).

> >> It does start to seem like a show stopper :-(
> >
> > I suppose that an srcu_read_lock_nmi() and srcu_read_unlock_nmi() could
> > be added, which would do atomic ops on sp->sda->srcu_lock_count. Not sure
> > whether this would be fast enough to be useful, but easy to provide:
> >
> > int __srcu_read_lock_nmi(struct srcu_struct *sp) /* UNTESTED. */
> > {
> > int idx;
> >
> > idx = READ_ONCE(sp->srcu_idx) & 0x1;
> > atomic_inc(&sp->sda->srcu_lock_count[idx]);
> > smp_mb__after_atomic(); /* B */ /* Avoid leaking critical section. */
> > return idx;
> > }
> >
> > void __srcu_read_unlock_nmi(struct srcu_struct *sp, int idx)
> > {
> > smp_mb__before_atomic(); /* C */ /* Avoid leaking critical section. */
> > atomic_inc(&sp->sda->srcu_unlock_count[idx]);
> > }
> >
> > With appropriate adjustments to also allow Tiny RCU to also work.
> >
> > Note that you have to use _nmi() everywhere, not just in NMI handlers.
> > In fact, the NMI handlers are the one place you -don't- need to use
> > _nmi(), strangely enough.
> >
> > Might be worth a try -- smp_mb__{before,after}_atomic() is a no-op on
> > some architectures, for example.
>
> Continuing Steve's question on regular interrupts, do we need to use
> this atomic_inc API for regular interrupts as well?

If NMIs use one srcu_struct and non-NMI uses another, the current
srcu_read_lock() and srcu_read_unlock() will work just fine. If any given
srcu_struct needs both NMI and non-NMI readers, then we really do need
__srcu_read_lock_nmi() and __srcu_read_unlock_nmi() for that srcu_struct.

Thanx, Paul