Re: [PATCH RFC v2 tip/core/rcu] Maintain special bits at bottom of ->dynticks counter

From: Paul E. McKenney
Date: Mon Feb 13 2017 - 12:01:12 EST


On Mon, Feb 13, 2017 at 01:21:15PM +0100, Peter Zijlstra wrote:
> On Thu, Feb 09, 2017 at 03:51:03PM -0800, Paul E. McKenney wrote:
> > Currently, IPIs are used to force other CPUs to invalidate their TLBs
> > in response to a kernel virtual-memory mapping change. This works, but
> > degrades both battery lifetime (for idle CPUs) and real-time response
> > (for nohz_full CPUs), and in addition results in unnecessary IPIs due to
> > the fact that CPUs executing in usermode are unaffected by stale kernel
> > mappings. It would be better to cause a CPU executing in usermode to
> > wait until it is entering kernel mode to do the flush, first to avoid
> > interrupting usemode tasks and second to handle multiple flush requests
> > with a single flush in the case of a long-running user task.
> >
> > This commit therefore reserves a bit at the bottom of the ->dynticks
> > counter, which is checked upon exit from extended quiescent states.
> > If it is set, it is cleared and then a new rcu_eqs_special_exit() macro is
> > invoked, which, if not supplied, is an empty single-pass do-while loop.
> > If this bottom bit is set on -entry- to an extended quiescent state,
> > then a WARN_ON_ONCE() triggers.
> >
> > This bottom bit may be set using a new rcu_eqs_special_set() function,
> > which returns true if the bit was set, or false if the CPU turned
> > out to not be in an extended quiescent state. Please note that this
> > function refuses to set the bit for a non-nohz_full CPU when that CPU
> > is executing in usermode because usermode execution is tracked by RCU
> > as a dyntick-idle extended quiescent state only for nohz_full CPUs.
> >
> > Changes since v1: Fix ordering of atomic_and() and the call to
> > rcu_eqs_special_exit() in rcu_dynticks_eqs_exit().
> >
> > Reported-by: Andy Lutomirski <luto@xxxxxxxxxxxxxx>
> > Signed-off-by: Paul E. McKenney <paulmck@xxxxxxxxxxxxxxxxxx>
>
> I think I've asked this before, but why does this live in the guts of
> RCU?
>
> Should we lift this state tracking stuff out and make RCU and
> NOHZ(_FULL) users of it, or doesn't that make sense (reason)?

The dyntick-idle stuff is pretty specific to RCU. And what precisely
would be helped by moving it?

But that was an excellent question, as it reminded me of RCU's
dyntick-idle's NMI handling, and I never did ask Andy if it was OK for
rcu_eqs_special_exit() to be invoked when exiting NMI handler, which would
currently happen. It would be easy for me to pass in a flag indicating
whether or not the call is in NMI context, if that is needed.

It is of course not possible to detect this at rcu_eqs_special_set()
time, because rcu_eqs_special_set() has no way of knowing that the next
event that pulls the remote CPU out of idle will be an NMI.

> In any case, small nit below:
>
>
> > + seq = atomic_add_return(RCU_DYNTICK_CTRL_CTR, &rdtp->dynticks);
> > + WARN_ON_ONCE(IS_ENABLED(CONFIG_RCU_EQS_DEBUG) &&
> > + !(seq & RCU_DYNTICK_CTRL_CTR));
> > + if (seq & RCU_DYNTICK_CTRL_MASK) {
> > + atomic_and(~RCU_DYNTICK_CTRL_MASK, &rdtp->dynticks);
> > + smp_mb__after_atomic(); /* _exit after clearing mask. */
> > + /* Prefer duplicate flushes to losing a flush. */
> > + rcu_eqs_special_exit();
> > + }
>
> we have atomic_andnot() for just these occasions :-)

I suppose that that could generate more efficient code on some
architectures. I have changed this.

> > +/*
> > + * Set the special (bottom) bit of the specified CPU so that it
> > + * will take special action (such as flushing its TLB) on the
> > + * next exit from an extended quiescent state. Returns true if
> > + * the bit was successfully set, or false if the CPU was not in
> > + * an extended quiescent state.
> > + */
> > +bool rcu_eqs_special_set(int cpu)
> > +{
> > + int old;
> > + int new;
> > + struct rcu_dynticks *rdtp = &per_cpu(rcu_dynticks, cpu);
> > +
> > + do {
> > + old = atomic_read(&rdtp->dynticks);
> > + if (old & RCU_DYNTICK_CTRL_CTR)
> > + return false;
> > + new = old | RCU_DYNTICK_CTRL_MASK;
> > + } while (atomic_cmpxchg(&rdtp->dynticks, old, new) != old);
> > + return true;
>
> Is that what we call atomic_fetch_or() ?

I don't think so. The above code takes an early exit if the next bit
up is set, which atomic_fetch_or() does not. If the CPU is not in
an extended quiescent state (old & RCU_DYNTICK_CTRL_CTR), then this
code returns false to indicate that TLB shootdown cannot wait.
So it is more like a very specific form of atomic_fetch_or_unless().

Thanx, Paul