Re: [patch V6 12/37] x86/entry: Provide idtentry_entry/exit_cond_rcu()

[Paul E. McKenney]

On Wed, May 20, 2020 at 03:15:31PM -0700, Paul E. McKenney wrote:
> On Wed, May 20, 2020 at 09:49:18PM +0200, Thomas Gleixner wrote:
> > "Paul E. McKenney" <paulmck@xxxxxxxxxx> writes:
> > > On Wed, May 20, 2020 at 09:51:17AM -0700, Andy Lutomirski wrote:
> > >> Paul, the major change here is that if an IRQ hits normal kernel code
> > >> (i.e. code where RCU is watching and we're not in an EQS), the IRQ
> > >> won't call rcu_irq_enter() and rcu_irq_exit().  Instead it will call
> > >> rcu_tickle() on entry and nothing on exit.  Does that cover all the
> > >> bases?
> > >
> > > From an RCU viewpoint, yes, give or take my concerns about someone
> > > putting rcu_tickle() on entry and rcu_irq_exit() on exit.  Perhaps
> > > I can bring some lockdep trickery to bear.
> > 
> > An surplus rcu_irq_exit() should already trigger alarms today.
> 
> Fair point!
> 
> > > But I must defer to Thomas and Peter on the non-RCU/non-NO_HZ_FULL
> > > portions of this.
> > 
> > I don't see a problem. Let me write that into actual testable patch
> > form.
> 
> Here is the RCU part, with my current best guess for the commit log.
> 
> Please note that this is on top of my -rcu stack, so some adjustment
> will likely be needed to pull it underneath Joel's series that removes
> the special-purpose bits at the bottom of the ->dynticks counter.
> 
> But a starting point, anyway.

Same patch, but with updated commit log based on IRC discussion
with Andy.

							Thanx, Paul

------------------------------------------------------------------------

commit 1771ea9fac5748d1424d9214c51b2f79cc1176b6
Author: Paul E. McKenney <paulmck@xxxxxxxxxx>
Date:   Wed May 20 15:03:07 2020 -0700

    rcu: Abstract out tickle_nohz_for_rcu() from rcu_nmi_enter()
    
    There will likely be exception handlers that can sleep, which rules
    out the usual approach of invoking rcu_nmi_enter() on entry and also
    rcu_nmi_exit() on all exit paths.  However, the alternative approach of
    just not calling anything can prevent RCU from coaxing quiescent states
    from nohz_full CPUs that are looping in the kernel:  RCU must instead
    IPI them explicitly.  It would be better to enable the scheduler tick
    on such CPUs to interact with RCU in a lighter-weight manner, and this
    enabling is one of the things that rcu_nmi_enter() currently does.
    
    What is needed is something that helps RCU coax quiescent states while
    not preventing subsequent sleeps.  This commit therefore splits out the
    nohz_full scheduler-tick enabling from the rest of the rcu_nmi_enter()
    logic into a new function named tickle_nohz_for_rcu().
    
    Suggested-by: Andy Lutomirski <luto@xxxxxxxxxx>
    Signed-off-by: Paul E. McKenney <paulmck@xxxxxxxxxx>

diff --git a/include/linux/hardirq.h b/include/linux/hardirq.h
index 621556e..d4be42a 100644
--- a/include/linux/hardirq.h
+++ b/include/linux/hardirq.h
@@ -14,6 +14,10 @@ extern bool synchronize_hardirq(unsigned int irq);
 
 #if defined(CONFIG_TINY_RCU)
 
+static inline void tickle_nohz_for_rcu(void)
+{
+}
+
 static inline void rcu_nmi_enter(void)
 {
 }
@@ -23,6 +27,7 @@ static inline void rcu_nmi_exit(void)
 }
 
 #else
+extern void tickle_nohz_for_rcu(void);
 extern void rcu_nmi_enter(void);
 extern void rcu_nmi_exit(void);
 #endif
diff --git a/kernel/rcu/tree.c b/kernel/rcu/tree.c
index 7812574..0a3cad4 100644
--- a/kernel/rcu/tree.c
+++ b/kernel/rcu/tree.c
@@ -806,6 +806,67 @@ void noinstr rcu_user_exit(void)
 #endif /* CONFIG_NO_HZ_FULL */
 
 /**
+ * tickle_nohz_for_rcu - Enable scheduler tick on CPU if RCU needs it.
+ *
+ * The scheduler tick is not normally enabled when CPUs enter the kernel
+ * from nohz_full userspace execution.  After all, nohz_full userspace
+ * execution is an RCU quiescent state and the time executing in the kernel
+ * is quite short.  Except of course when it isn't.  And it is not hard to
+ * cause a large system to spend tens of seconds or even minutes looping
+ * in the kernel, which can cause a number of problems, include RCU CPU
+ * stall warnings.
+ *
+ * Therefore, if a nohz_full CPU fails to report a quiescent state
+ * in a timely manner, the RCU grace-period kthread sets that CPU's
+ * ->rcu_urgent_qs flag with the expectation that the next interrupt or
+ * exception will invoke this function, which will turn on the scheduler
+ * tick, which will enable RCU to detect that CPU's quiescent states,
+ * for example, due to cond_resched() calls in CONFIG_PREEMPT=n kernels.
+ * The tick will be disabled once a quiescent state is reported for
+ * this CPU.
+ *
+ * Of course, in carefully tuned systems, there might never be an
+ * interrupt or exception.  In that case, the RCU grace-period kthread
+ * will eventually cause one to happen.  However, in less carefully
+ * controlled environments, this function allows RCU to get what it
+ * needs without creating otherwise useless interruptions.
+ */
+noinstr void tickle_nohz_for_rcu(void)
+{
+	struct rcu_data *rdp = this_cpu_ptr(&rcu_data);
+
+	if (in_nmi())
+		return; // Enabling tick is unsafe in NMI handlers.
+	RCU_LOCKDEP_WARN(rcu_dynticks_curr_cpu_in_eqs(),
+			 "Illegal tickle_nohz_for_rcu from extended quiescent state");
+	instrumentation_begin();
+	if (!tick_nohz_full_cpu(rdp->cpu) ||
+	    !READ_ONCE(rdp->rcu_urgent_qs) ||
+	    READ_ONCE(rdp->rcu_forced_tick)) {
+		// RCU doesn't need nohz_full help from this CPU, or it is
+		// already getting that help.
+		instrumentation_end();
+		return;
+	}
+
+	// We get here only when not in an extended quiescent state and
+	// from interrupts (as opposed to NMIs).  Therefore, (1) RCU is
+	// already watching and (2) The fact that we are in an interrupt
+	// handler and that the rcu_node lock is an irq-disabled lock
+	// prevents self-deadlock.  So we can safely recheck under the lock.
+	// Note that the nohz_full state currently cannot change.
+	raw_spin_lock_rcu_node(rdp->mynode);
+	if (rdp->rcu_urgent_qs && !rdp->rcu_forced_tick) {
+		// A nohz_full CPU is in the kernel and RCU needs a
+		// quiescent state.  Turn on the tick!
+		WRITE_ONCE(rdp->rcu_forced_tick, true);
+		tick_dep_set_cpu(rdp->cpu, TICK_DEP_BIT_RCU);
+	}
+	raw_spin_unlock_rcu_node(rdp->mynode);
+	instrumentation_end();
+}
+
+/**
  * rcu_nmi_enter - inform RCU of entry to NMI context
  * @irq: Is this call from rcu_irq_enter?
  *
@@ -835,7 +896,9 @@ noinstr void rcu_nmi_enter(void)
 	 * is if the interrupt arrived in kernel mode; in this case we would
 	 * be the outermost interrupt but still increment by 2 which is Ok.
 	 */
-	if (rcu_dynticks_curr_cpu_in_eqs()) {
+	if (!rcu_dynticks_curr_cpu_in_eqs()) {
+		tickle_nohz_for_rcu();
+	} else {
 
 		if (!in_nmi())
 			rcu_dynticks_task_exit();
@@ -851,28 +914,6 @@ noinstr void rcu_nmi_enter(void)
 		}
 
 		incby = 1;
-	} else if (!in_nmi()) {
-		instrumentation_begin();
-		if (tick_nohz_full_cpu(rdp->cpu) &&
-		    READ_ONCE(rdp->rcu_urgent_qs) &&
-		    !READ_ONCE(rdp->rcu_forced_tick)) {
-			// We get here only if we had already exited the
-			// extended quiescent state and this was an
-			// interrupt (not an NMI).  Therefore, (1) RCU is
-			// already watching and (2) The fact that we are in
-			// an interrupt handler and that the rcu_node lock
-			// is an irq-disabled lock prevents self-deadlock.
-			// So we can safely recheck under the lock.
-			raw_spin_lock_rcu_node(rdp->mynode);
-			if (rdp->rcu_urgent_qs && !rdp->rcu_forced_tick) {
-				// A nohz_full CPU is in the kernel and RCU
-				// needs a quiescent state.  Turn on the tick!
-				WRITE_ONCE(rdp->rcu_forced_tick, true);
-				tick_dep_set_cpu(rdp->cpu, TICK_DEP_BIT_RCU);
-			}
-			raw_spin_unlock_rcu_node(rdp->mynode);
-		}
-		instrumentation_end();
 	}
 	instrumentation_begin();
 	trace_rcu_dyntick(incby == 1 ? TPS("End") : TPS("StillNonIdle"),