Re: [RFC, PATCH -rt] NMI-safe mb- and atomic-free RT RCU

On Thu, 27 Jul 2006, Paul E. McKenney wrote:

On Thu, Jul 27, 2006 at 12:00:13PM +0100, Esben Nielsen wrote:

Hi Paul,
 Thanks, for entering a discussion of my idea! Even though you are busy
and are critical towards my idea, you take your time to answer! Thanks.

I appreciate the time you took to write down your ideas, and for your
taking my comments in the spirit intended.

Well, I just think it is fun :-)

On Wed, 26 Jul 2006, Paul E. McKenney wrote:

On Thu, Jul 27, 2006 at 02:39:07AM +0100, Esben Nielsen wrote:

On Tue, 25 Jul 2006, Paul E. McKenney wrote:

Not for inclusion, should be viewed with great suspicion.

This patch provides an NMI-safe realtime RCU.  Hopefully, Mathieu can
make use of synchronize_sched() instead, since I have not yet figured
out a way to make this NMI-safe and still get rid of the interrupt
disabling.  ;-)

						Thanx, Paul

I must say I don't understand all this. It looks very complicated. Is it
really needed?

I have been thinking about the following design:

void rcu_read_lock()
{
	if (!in_interrupt())
		current->rcu_read_lock_count++;
}
void rcu_read_unlock()
{
	if (!in_interrupt())
		current->rcu_read_lock_count--;
}

Somewhere in schedule():

	rq->rcu_read_lock_count += prev->rcu_read_lock_count;
	if (!rq->rcu_read_lock_count)
		forward_waiting_rcu_jobs();
	rq->rcu_read_lock_count -= next->rcu_read_lock_count;

So rq->rcu_read_lock_count contains the sum of the counts of all
tasks that were scheduled away from this CPU.

What happens in face of the following sequence of events?
Assume that all tasks stay on CPU 0 for the moment.

o	Task A does rcu_read_lock().

o	Task A is preempted.  rq->rcu_read_lock_count is nonzero.

o	Task B runs and does rcu_read_lock().

o	Task B is preempted (perhaps because it dropped the lock
	that was causing it to have high priority).  Regardless of
	the reason for the preemption, rq->rcu_read_lock_count is
	nonzero.

o	Task C runs and does rcu_read_lock().

o	Task C is preempted.  rq->rcu_read_lock_count is nonzero.

o	Task A runs again, and does rcu_read_unlock().

o	Task A is preempted.  rq->rcu_read_lock_count is nonzero.

o	Task D runs and does rcu_read_lock().

o	Task D is preempted.  rq->rcu_read_lock_count is nonzero.

And so on.  As long as at least one of the preempted tasks is in an
RCU critical section, you never do your forward_waiting_rcu_jobs(),
and the grace period goes on forever.  Or at least until you OOM the
machine.

So what am I missing here?

The boosting idea below. Then tasks A-D must be RT tasks for this to
happen. And the machine must anyway run out of RT tasks or it will
effectively lock up.

I agree that boosting would drive any particular task out of its
RCU read-side critical section, but I do not agree that you have
identified a valid induction argument that proves that we run out
of tasks (yet, anyway).  For one thing, tasks acquiring locks in
the RCU read-side critical sections could be priority boosted,
so that a small set of tasks could rise and fall in priority as
RT tasks attempted to acquire locks held by this small set of
tasks.  The RT tasks need not be running on the same CPU that the
non-realtime "victim" tasks running in RCU read-side critical sections,
therefore, the RT tasks need not be preempting these non-realtime
tasks.

To be more specific, imagine a four-CPU system with three RT tasks
running on three of the four CPUs and a large number of non-realtime
tasks running on all four CPUs.  Consider the code in __d_lookup(), which
acquires dentry->d_lock under rcu_read_lock().  Imagine a situation
where the three RT tasks are repeatedly traversing pathnames with a
long common path prefix, with this same prefix being traversed by the
non-realtime tasks.  This could result in the non-realtime tasks being
repeatedly priority-boosted and preempted (preempted both by each
other and by RT tasks).

Over to you!

The PI code will never lower the priority lower than task->normal_prio.

So if you somewhere in schedule() set normal_prio to something high PI will not destroy it.

I could imagine introducing a pair of counters per runqueue, but then
we end up with the same issues with counter-flip that we have now.

Another question -- what happens if a given CPU stays idle?  Wouldn't
the callbacks then just start piling up on that CPU?

How can a CPU stay idle? There is a tick every 2.5 ms. Even without that
the previous CPU can make it schedule if it sees the jobs piling up. Or if
that is considered too expensive, it can take over and forward the
jobs to the next CPU.

Tickless idle, prized by embedded systems.  Here an idle CPU gets shut
down if there is nothing in particular for it to do.  Classic RCU does a
dance with CPUs entering tickless idle state in order to avoid expecting
those CPUs to participate in future grace periods.

This had its challenges for the bitmask-based Classic RCU, and would
have a few additional challenges for your circular linked-list setup.

Can't the previous CPU just see that when it wants to send the RCU-jobs along to it and send it to the next CPU instead? Or when a CPU goes idle it can take itself out of the circle to avoid the extra load on the previous CPU.

Current -rt RCU avoids this problem by dint of being too stupid to
need to know what CPUs are alive or not.  My recent patch needs something
similar to the tickless-idle change to Classic RCU.

Now what should forward_waiting_rcu_jobs() do?

I imagine a circular datastructur of all the CPUs. When a call_rcu() is
run on a CPU it is first added a list of jobs for that CPU. When
forward_waiting_rcu_jobs() is called all the pending jobs are forwarded to
the next CPU. The next CPU will bring it along the next CPU in the circle
along with it's own jobs. When jobs hit the original CPU they will be
executed. Or rather, when the CPU just before calls
forward_waiting_rcu_jobs(), it sends the jobs belonging to the next CPU to
the RCU-task of the next CPU, where they will be executed, instead of to
the scheduler (runqueue) of the next CPU, where it will just be send out
on
a
new roundtrip along the circle.

If you use a structure like the plist then the forwarding procedure can be
done in O(number of online CPUs) time worst case - much less in the usual
case where the lists are almost empty.

Now the problem is: What happens if a task in a rcu read-side lock is
migrated? Then the rcu_read_lock_count on the source CPU will stay in plus
while on the target CPU it will go in minus. This ought to be simply
fixeable by adding task->rcu_read_lock_count to the target runqueue before
migrating and subtracting it from the old runqueue after migrating. But
there is another problem: RCU-jobs refering to data used by the task being
migrated might have been forwarded from the target CPU. Thus the migration
task have to go back along the circle of CPUs and move all the relevant
RCU-jobs back to the target CPU to be forwarded again. This is also doable
in
number of CPUs between source and target times O(<number of online CPUs>)
(see above) time.

So if I have the right (or wrong) pattern of task migrations, the RCU
callbacks never get to their originating CPU?

In principle, yes. But if the machine starts to migrate tasks that wildly
it wont get any work done anyway, because all it's time is done doing
migration.

Since you are waiting for a context switch to move the RCU callbacks,
the migration does not have to be all that wild.  In an eight-CPU system,
if I migrate a preempted task a couple of times in the course of the
eight context switches required for the callbacks to circle the loop
of CPUs, we are in trouble, right?

Could be. The other idea I came up with below doesn't have that problem.

Alternatively, if the task residing in the RCU read-side critical section
is forwarded around the loop of CPUs, callbacks circulating around this
loop might execute before the RCU read-side critical section completes.

That is why some of the callbacks (those which has parsed the target CPU
but not yet the source CPU) have to be moved back to the target CPU.

Yep.  Leaving them to follow their planned circuit indeed does not cut it.

I just came up with an even simpler solution:
Delay the subtraction of the task->rcu_read_lock_count from
srcrq->rcu_read_lock_count until the task calls rcu_read_unlock(). That
can be done by flagging the task (do task->rcu_read_lock_count |=
0x80000000) and do a simple
	if (unlickely(current->rcu_read_lock_count == 0x80000000))
		fix_rcu_read_lock_count_on_old_cpu();
in rcu_read_unlock(). Now the task can be migrated again before calloing
fix_rcu_read_lock_count_on_old_cpu(). The relevant RCU jobs still can't
get past the original CPU before the task have called
fix_rcu_read_lock_count_on_old_cpu(), so all subsequent migrations can just
do the count down on the intermediate CPUs right away.

But now you need atomic instructions in all rcu_read_lock() and
rcu_read_unlock() invocations to handle the possibility that some
migrated task will decrement the counter at the same time we are
incrementing it?  I cannot tell for sure because you have not supplied
sample code for fix_rcu_read_lock_count_on_old_cpu().

Now the atomic part is only kicking in if the task was actually migrated while under RCU readside section so it is rare. But consider this code:

void fix_rcu_read_lock_count_on_old_cpu()
{
	int cpuid = current->old_cpuid;
	int count = current->old_rcu_read_lock_count;

	runqueue_t *rq = cpu_rq(cpuid);
        spin_lock_irqsave(&rq->lock, flags);
	rq->rcu_read_lock_count -= count;
        spin_unlock_irqrestore(&rq->lock, flags);
}

Then there can be something about if the CPU was taken out meanwhile. If the rq->rcu_read_lock is 0 after this, then it can go completely out of the RCU CPU circle now (see below).

To avoid a task in a read-side lock being starved for too long the
following line can be added to normal_prio():
 if (p->rcu_read_lock_count)
	p->prio = MAX_RT_PRIO;

But doesn't this have the same impact on latency as disabling preemption
in rcu_read_lock() and then re-enabling it in rcu_read_unlock()?

No, RT tasks can still preempt the RCU read side lock. But SCHED_OTHER and
SCHED_BATCH can't. You can also the RCU read side boosting prioritiy
dynamic and let the system adjust it or just let the admin adjust it.

Fair enough -- I misread MAX_RT_PRIO as MAX_PRIO.

This approach I can get behind -- my thought has been to boost to
either MAX_RT_PRIO or MAX_RT_PRIO-1 when preempt_schedule() sees that
it is preempting an RCU read-side critical section.

So I agree with you on at least one point!  ;-)

A possible elaboration would be to keep a linked list of tasks preempted
in their RCU read-side critical sections so that they can be further
boosted to the highest possible priority (numerical value of zero,
not sure what the proper symbol is) if the grace period takes too many
jiffies to complete.  Another piece is priority boosting when blocking
on a mutex from within an RCU read-side critical section.

Also, doesn't this circular data structure need to handle CPU hotplug?

Ofcourse. I don't know about hotplug though. But it sounds simple to
migrate the tasks away, take the CPU out of the circle and then forward
the last RCU jobs from that CPU.

Doing it efficiently is the difficulty, particularly for tickless-idle
systems where CPUs need to be added and removed on a regular basis.
Also, what locking design would you use in order to avoid deadlock?
There is a hotplug mutex, but seems like you might need to acquire some
number of rq mutexes as well.

I think it can be done very effectively. When taking a CPU out: If, after migrating all the tasks off the CPU, the rq->rcu_read_lock_count is 0, alter the previous CPUs next CPU to this CPUs next CPU. Then forward the pending RCU jobs. If some of the migrated task where in an RCU read side section rq->rcu_read_lock_count > 0. Then the CPU has to be keeped in the RCU circle until all are done. Either an event can be set up or the previous CPU can check on it everytime it tries to forward jobs to the CPU. When putting the CPU in again: Simply add it to the circle when it is up running.

I don't have time to code this nor a SMP machine to test it on. But I can
give the idea to you anyways in the hope you might code it :-)

I am beginning to think that it will not be at all simple by the time I
code up all the required fixups.  Or am I missing something?

Ofcourse, implementing something is always a lot harder than writing the
idea down. Anyway, we already worked out some of the hardest details :-)

I certainly agree with your first sentence.  On your second sentence,
I would s/worked out/mentioned/  ;-)

Another approach I am looking at does not permit rcu_read_lock() in
NMI/SMI/hardirq, but is much simpler.  Its downside is that it cannot
serve as common code between CONFIG_PREEMPT_RT and CONFIG_PREEMPT.

Iack :-( That is not good. :-(

Esben

							Thanx, Paul

Signed-off-by: Paul E. McKenney <[email protected]> (but please don't
apply)

include/linux/sched.h |    3
kernel/rcupreempt.c   |  577
++++++++++++++++++++++++++++++++++++++++----------
2 files changed, 470 insertions(+), 110 deletions(-)

diff -urpNa -X dontdiff linux-2.6.17-rt7/include/linux/sched.h
linux-2.6.17-rt7-norrupt/include/linux/sched.h
--- linux-2.6.17-rt7/include/linux/sched.h	2006-07-19
01:43:09.000000000 -0700
+++ linux-2.6.17-rt7-norrupt/include/linux/sched.h	2006-07-22
15:44:36.000000000 -0700
@@ -868,8 +868,7 @@ struct task_struct {

#ifdef CONFIG_PREEMPT_RCU
	int rcu_read_lock_nesting;
-	atomic_t *rcu_flipctr1;
-	atomic_t *rcu_flipctr2;
+	int rcu_flipctr_idx;
#endif
#ifdef CONFIG_SCHEDSTATS
	struct sched_info sched_info;
diff -urpNa -X dontdiff linux-2.6.17-rt7/kernel/rcupreempt.c
linux-2.6.17-rt7-norrupt/kernel/rcupreempt.c
--- linux-2.6.17-rt7/kernel/rcupreempt.c	2006-07-19
01:43:09.000000000 -0700
+++ linux-2.6.17-rt7-norrupt/kernel/rcupreempt.c	2006-07-22
20:21:46.000000000 -0700
@@ -15,11 +15,13 @@
* along with this program; if not, write to the Free Software
* Foundation, Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307,
USA.
*
- * Copyright (C) IBM Corporation, 2001
+ * Copyright (C) IBM Corporation, 2006
*
* Authors: Paul E. McKenney <[email protected]>
*		With thanks to Esben Nielsen, Bill Huey, and Ingo Molnar
- *		for pushing me away from locks and towards counters.
+ *		for pushing me away from locks and towards counters, and
+ *		to Suparna Bhattacharya for pushing me completely away
+ *		from atomic instructions on the read side.
*
* Papers:  http://www.rdrop.com/users/paulmck/RCU
*
@@ -73,12 +75,20 @@ struct rcu_data {
	long		n_done_remove;
	atomic_t	n_done_invoked;
	long		n_rcu_check_callbacks;
-	atomic_t	n_rcu_try_flip1;
-	long		n_rcu_try_flip2;
-	long		n_rcu_try_flip3;
+	atomic_t	n_rcu_try_flip_1;
	atomic_t	n_rcu_try_flip_e1;
-	long		n_rcu_try_flip_e2;
-	long		n_rcu_try_flip_e3;
+	long		n_rcu_try_flip_i1;
+	long		n_rcu_try_flip_ie1;
+	long		n_rcu_try_flip_g1;
+	long		n_rcu_try_flip_a1;
+	long		n_rcu_try_flip_ae1;
+	long		n_rcu_try_flip_a2;
+	long		n_rcu_try_flip_z1;
+	long		n_rcu_try_flip_ze1;
+	long		n_rcu_try_flip_z2;
+	long		n_rcu_try_flip_m1;
+	long		n_rcu_try_flip_me1;
+	long		n_rcu_try_flip_m2;
#endif /* #ifdef CONFIG_RCU_STATS */
};
struct rcu_ctrlblk {
@@ -90,8 +100,51 @@ static struct rcu_ctrlblk rcu_ctrlblk =
	.fliplock = RAW_SPIN_LOCK_UNLOCKED,
	.completed = 0,
};
-static DEFINE_PER_CPU(atomic_t [2], rcu_flipctr) =
-	{ ATOMIC_INIT(0), ATOMIC_INIT(0) };
+static DEFINE_PER_CPU(int [2], rcu_flipctr) = { 0, 0 };
+
+/*
+ * States for rcu_try_flip() and friends.
+ */
+
+enum rcu_try_flip_states {
+	rcu_try_flip_idle_state,	/* "I" */
+	rcu_try_flip_in_gp_state,	/* "G" */
+	rcu_try_flip_waitack_state, 	/* "A" */
+	rcu_try_flip_waitzero_state,	/* "Z" */
+	rcu_try_flip_waitmb_state	/* "M" */
+};
+static enum rcu_try_flip_states rcu_try_flip_state =
rcu_try_flip_idle_state;
+#ifdef CONFIG_RCU_STATS
+static char *rcu_try_flip_state_names[] =
+	{ "idle", "gp", "waitack", "waitzero", "waitmb" };
+#endif /* #ifdef CONFIG_RCU_STATS */
+
+/*
+ * Enum and per-CPU flag to determine when each CPU has seen
+ * the most recent counter flip.
+ */
+
+enum rcu_flip_flag_values {
+	rcu_flip_seen,		/* Steady/initial state, last flip seen. */
+				/* Only GP detector can update. */
+	rcu_flipped		/* Flip just completed, need confirmation. */
+				/* Only corresponding CPU can update. */
+};
+static DEFINE_PER_CPU(enum rcu_flip_flag_values, rcu_flip_flag) =
rcu_flip_seen;
+
+/*
+ * Enum and per-CPU flag to determine when each CPU has executed the
+ * needed memory barrier to fence in memory references from its last RCU
+ * read-side critical section in the just-completed grace period.
+ */
+
+enum rcu_mb_flag_values {
+	rcu_mb_done,		/* Steady/initial state, no mb()s required.
*/
+				/* Only GP detector can update. */
+	rcu_mb_needed		/* Flip just completed, need an mb(). */
+				/* Only corresponding CPU can update. */
+};
+static DEFINE_PER_CPU(enum rcu_mb_flag_values, rcu_mb_flag) =
rcu_mb_done;

/*
* Return the number of RCU batches processed thus far.  Useful
@@ -105,93 +158,182 @@ long rcu_batches_completed(void)
void
rcu_read_lock(void)
{
-	int flipctr;
+	int idx;
+	int nesting;
	unsigned long oldirq;

-	local_irq_save(oldirq);
-
+	raw_local_irq_save(oldirq);
	trace_special(current->rcu_read_lock_nesting,
		      (unsigned long) current->rcu_flipctr1,
		      rcu_ctrlblk.completed);
-	if (current->rcu_read_lock_nesting++ == 0) {
+
+	nesting = current->rcu_read_lock_nesting;
+
+	/*
+	 * Any rcu_read_lock()s called from NMI handlers
+	 * at any point must have matching rcu_read_unlock()
+	 * calls in that same handler, so we will not see the
+	 * value of current->rcu_read_lock_nesting change.
+	 */
+
+	if (nesting != 0) {

		/*
-		 * Outermost nesting of rcu_read_lock(), so atomically
+		 * There is an enclosing rcu_read_lock(), so all we
+		 * need to do is to increment the counter.
+		 */
+
+		current->rcu_read_lock_nesting = nesting + 1;
+
+	} else {
+
+		/*
+		 * Outermost nesting of rcu_read_lock(), so we must
		 * increment the current counter for the current CPU.
+		 * This must be done carefully, because NMIs can
+		 * occur at any point in this code, and any rcu_read_lock()
+		 * and rcu_read_unlock() pairs in the NMI handlers
+		 * must interact non-destructively with this code.
+		 * Lots of barrier() calls, and -very- careful ordering.
+		 *
+		 * Changes to this code, including this one, must be
+		 * inspected, validated, and tested extremely carefully!!!
		 */

-		flipctr = rcu_ctrlblk.completed & 0x1;
+		/*
+		 * First, pick up the index.  Enforce ordering for
+		 * both compilers and for DEC Alpha.
+		 */
+
+		idx = rcu_ctrlblk.completed & 0x1;
		smp_read_barrier_depends();
-		current->rcu_flipctr1 =
&(__get_cpu_var(rcu_flipctr)[flipctr]);
-		/* Can optimize to non-atomic on fastpath, but start simple.
*/
-		atomic_inc(current->rcu_flipctr1);
-		smp_mb__after_atomic_inc();  /* might optimize out... */
-		if (unlikely(flipctr != (rcu_ctrlblk.completed & 0x1))) {
-
-			/*
-			 * We raced with grace-period processing (flip).
-			 * Although we cannot be preempted here, there
-			 * could be interrupts, ECC errors and the like,
-			 * so just nail down both sides of the rcu_flipctr
-			 * array for the duration of our RCU read-side
-			 * critical section, preventing a second flip
-			 * from racing with us.  At some point, it would
-			 * be safe to decrement one of the counters, but
-			 * we have no way of knowing when that would be.
-			 * So just decrement them both in rcu_read_unlock().
-			 */
-
-			current->rcu_flipctr2 =
-				&(__get_cpu_var(rcu_flipctr)[!flipctr]);
-			/* Can again optimize to non-atomic on fastpath. */
-			atomic_inc(current->rcu_flipctr2);
-			smp_mb__after_atomic_inc();  /* might optimize
out... */
-		}
+		barrier();
+
+		/*
+		 * Increment the per-CPU counter. NMI handlers
+		 * might increment it as well, but they had better
+		 * properly nest their rcu_read_lock()/rcu_read_unlock()
+		 * pairs so that the value is restored before the handler
+		 * returns to us.  Enforce ordering for compilers.
+		 */
+
+		__get_cpu_var(rcu_flipctr)[idx]++;
+		barrier();
+
+		/*
+		 * It is now safe to increment the task's nesting count.
+		 * NMIs that occur after this statement will route
+		 * their rcu_read_lock() calls through the "then" clause
+		 * of this "if" statement, and thus will no longer come
+		 * through this path.  Enforce ordering for compilers.
+		 */
+
+		current->rcu_read_lock_nesting = nesting + 1;
+		barrier();
+
+		/*
+		 * It is now safe to store the index into our task
+		 * structure.  Doing so earlier would have resulted
+		 * in it getting clobbered by NMI handlers.
+		 */
+
+		current->rcu_flipctr_idx = idx;
	}
+
	trace_special((unsigned long) current->rcu_flipctr1,
		      (unsigned long) current->rcu_flipctr2,
		      rcu_ctrlblk.completed);
-	local_irq_restore(oldirq);
+	raw_local_irq_restore(oldirq);
}

void
rcu_read_unlock(void)
{
+	int idx;
+	int nesting;
	unsigned long oldirq;

-	local_irq_save(oldirq);
+	raw_local_irq_save(oldirq);
	trace_special((unsigned long) current->rcu_flipctr1,
		      (unsigned long) current->rcu_flipctr2,
		      current->rcu_read_lock_nesting);
-	if (--current->rcu_read_lock_nesting == 0) {
+
+	nesting = current->rcu_read_lock_nesting;
+
+	/*
+	 * Any rcu_read_lock()s called from NMI handlers
+	 * at any point must have matching rcu_read_unlock()
+	 * calls in that same handler, so we will not see the
+	 * value of current->rcu_read_lock_nesting change.
+	 */
+
+	if (nesting > 1) {

		/*
-		 * Just atomically decrement whatever we incremented.
-		 * Might later want to awaken some task waiting for the
-		 * grace period to complete, but keep it simple for the
-		 * moment.
+		 * There is an enclosing rcu_read_lock(), so all we
+		 * need to do is to decrement the counter.
		 */

-		smp_mb__before_atomic_dec();
-		atomic_dec(current->rcu_flipctr1);
-		current->rcu_flipctr1 = NULL;
-		if (unlikely(current->rcu_flipctr2 != NULL)) {
-			atomic_dec(current->rcu_flipctr2);
-			current->rcu_flipctr2 = NULL;
-		}
+		current->rcu_read_lock_nesting = nesting - 1;
+
+	} else {
+
+		/*
+		 * Outermost nesting of rcu_read_unlock(), so we must
+		 * decrement the current counter for the current CPU.
+		 * This must be done carefully, because NMIs can
+		 * occur at any point in this code, and any rcu_read_lock()
+		 * and rcu_read_unlock() pairs in the NMI handlers
+		 * must interact non-destructively with this code.
+		 * Lots of barrier() calls, and -very- careful ordering.
+		 *
+		 * Changes to this code, including this one, must be
+		 * inspected, validated, and tested extremely carefully!!!
+		 */
+
+		/*
+		 * First, pick up the index.  Enforce ordering for
+		 * both compilers and for DEC Alpha.
+		 */
+
+		idx = current->rcu_flipctr_idx;
+		smp_read_barrier_depends();
+		barrier();
+
+		/*
+		 * It is now safe to decrement the task's nesting count.
+		 * NMIs that occur after this statement will route
+		 * their rcu_read_lock() calls through this "else" clause
+		 * of this "if" statement, and thus will start incrementing
+		 * the per-CPU counter on their own.  Enforce ordering for
+		 * compilers.
+		 */
+
+		current->rcu_read_lock_nesting = nesting - 1;
+		barrier();
+
+		/*
+		 * Decrement the per-CPU counter. NMI handlers
+		 * might increment it as well, but they had better
+		 * properly nest their rcu_read_lock()/rcu_read_unlock()
+		 * pairs so that the value is restored before the handler
+		 * returns to us.
+		 */
+
+		__get_cpu_var(rcu_flipctr)[idx]--;
	}

	trace_special((unsigned long)current->rcu_flipctr1,
		      (unsigned long) current->rcu_flipctr2,
		      current->rcu_read_lock_nesting);
-	local_irq_restore(oldirq);
+	raw_local_irq_restore(oldirq);
}

static void
__rcu_advance_callbacks(void)
{

-	if (rcu_data.completed != rcu_ctrlblk.completed) {
+	if ((rcu_data.completed >> 1) != (rcu_ctrlblk.completed >> 1)) {
		if (rcu_data.waitlist != NULL) {
			*rcu_data.donetail = rcu_data.waitlist;
			rcu_data.donetail = rcu_data.waittail;
@@ -216,13 +358,186 @@ __rcu_advance_callbacks(void)
			rcu_data.waittail = &rcu_data.waitlist;
		}
		rcu_data.completed = rcu_ctrlblk.completed;
+	} else if (rcu_data.completed != rcu_ctrlblk.completed) {
+		rcu_data.completed = rcu_ctrlblk.completed;
+	}
+}
+
+/*
+ * Get here when RCU is idle.  Decide whether we need to
+ * move out of idle state, and return zero if so.
+ * "Straightforward" approach for the moment, might later
+ * use callback-list lengths, grace-period duration, or
+ * some such to determine when to exit idle state.
+ * Might also need a pre-idle test that does not acquire
+ * the lock, but let's get the simple case working first...
+ */
+
+static int
+rcu_try_flip_idle(int flipctr)
+{
+#ifdef CONFIG_RCU_STATS
+	rcu_data.n_rcu_try_flip_i1++;
+#endif /* #ifdef CONFIG_RCU_STATS */
+
+	if (!rcu_pending(smp_processor_id())) {
+#ifdef CONFIG_RCU_STATS
+		rcu_data.n_rcu_try_flip_ie1++;
+#endif /* #ifdef CONFIG_RCU_STATS */
+		return 1;
	}
+	return 0;
+}
+
+/*
+ * Flip processing up to and including the flip, as well as
+ * telling CPUs to acknowledge the flip.
+ */
+
+static int
+rcu_try_flip_in_gp(int flipctr)
+{
+	int cpu;
+
+#ifdef CONFIG_RCU_STATS
+	rcu_data.n_rcu_try_flip_g1++;
+#endif /* #ifdef CONFIG_RCU_STATS */
+
+	/*
+	 * Do the flip.
+	 */
+
+	rcu_ctrlblk.completed++;  /* stands in for rcu_try_flip_g2 */
+
+	/*
+	 * Need a memory barrier so that other CPUs see the new
+	 * counter value before they see the subsequent change of all
+	 * the rcu_flip_flag instances to rcu_flipped.
+	 */
+
+	smp_mb();
+
+	/* Now ask each CPU for acknowledgement of the flip. */
+
+	for_each_cpu(cpu)
+		per_cpu(rcu_flip_flag, cpu) = rcu_flipped;
+
+	return 0;
+}
+
+/*
+ * Wait for CPUs to acknowledge the flip.
+ */
+
+static int
+rcu_try_flip_waitack(int flipctr)
+{
+	int cpu;
+
+#ifdef CONFIG_RCU_STATS
+	rcu_data.n_rcu_try_flip_a1++;
+#endif /* #ifdef CONFIG_RCU_STATS */
+
+	for_each_cpu(cpu)
+		if (per_cpu(rcu_flip_flag, cpu) != rcu_flip_seen) {
+#ifdef CONFIG_RCU_STATS
+			rcu_data.n_rcu_try_flip_ae1++;
+#endif /* #ifdef CONFIG_RCU_STATS */
+			return 1;
+		}
+
+	/*
+	 * Make sure our checks above don't bleed into subsequent
+	 * waiting for the sum of the counters to reach zero.
+	 */
+
+	smp_mb();
+
+#ifdef CONFIG_RCU_STATS
+	rcu_data.n_rcu_try_flip_a2++;
+#endif /* #ifdef CONFIG_RCU_STATS */
+
+	return 0;
+}
+
+/*
+ * Wait for collective ``last'' counter to reach zero,
+ * then tell all CPUs to do an end-of-grace-period memory barrier.
+ */
+
+static int
+rcu_try_flip_waitzero(int flipctr)
+{
+	int cpu;
+	int lastidx = !(flipctr & 0x1);
+	int sum = 0;
+
+#ifdef CONFIG_RCU_STATS
+	rcu_data.n_rcu_try_flip_z1++;
+#endif /* #ifdef CONFIG_RCU_STATS */
+
+	/* Check to see if the sum of the "last" counters is zero. */
+
+	for_each_cpu(cpu)
+		sum += per_cpu(rcu_flipctr, cpu)[lastidx];
+	if (sum != 0) {
+#ifdef CONFIG_RCU_STATS
+		rcu_data.n_rcu_try_flip_ze1++;
+#endif /* #ifdef CONFIG_RCU_STATS */
+		return 1;
+	}
+
+#ifdef CONFIG_RCU_STATS
+	rcu_data.n_rcu_try_flip_z2++;
+#endif /* #ifdef CONFIG_RCU_STATS */
+
+	/* Make sure we don't call for memory barriers before we see zero. */
+
+	smp_mb();
+
+	/* Call for a memory barrier from each CPU. */
+
+	for_each_cpu(cpu)
+		per_cpu(rcu_mb_flag, cpu) = rcu_mb_needed;
+
+	return 0;
+}
+
+/*
+ * Wait for all CPUs to do their end-of-grace-period memory barrier.
+ * Return 0 once all CPUs have done so.
+ */
+
+static int
+rcu_try_flip_waitmb(int flipctr)
+{
+	int cpu;
+
+#ifdef CONFIG_RCU_STATS
+	rcu_data.n_rcu_try_flip_m1++;
+#endif /* #ifdef CONFIG_RCU_STATS */
+
+	for_each_cpu(cpu)
+		if (per_cpu(rcu_mb_flag, cpu) != rcu_mb_done) {
+#ifdef CONFIG_RCU_STATS
+			rcu_data.n_rcu_try_flip_me1++;
+#endif /* #ifdef CONFIG_RCU_STATS */
+			return 1;
+		}
+
+	smp_mb(); /* Ensure that the above checks precede any following
flip. */
+
+#ifdef CONFIG_RCU_STATS
+	rcu_data.n_rcu_try_flip_m2++;
+#endif /* #ifdef CONFIG_RCU_STATS */
+
+	return 0;
}

/*
* Attempt a single flip of the counters.  Remember, a single flip does
* -not- constitute a grace period.  Instead, the interval between
- * a pair of consecutive flips is a grace period.
+ * at least three consecutive flips is a grace period.
*
* If anyone is nuts enough to run this CONFIG_PREEMPT_RCU implementation
* on a large SMP, they might want to use a hierarchical organization of
@@ -231,13 +546,11 @@ __rcu_advance_callbacks(void)
static void
rcu_try_flip(void)
{
-	int cpu;
	long flipctr;
	unsigned long oldirq;

-	flipctr = rcu_ctrlblk.completed;
#ifdef CONFIG_RCU_STATS
-	atomic_inc(&rcu_data.n_rcu_try_flip1);
+	atomic_inc(&rcu_data.n_rcu_try_flip_1);
#endif /* #ifdef CONFIG_RCU_STATS */
	if (unlikely(!spin_trylock_irqsave(&rcu_ctrlblk.fliplock, oldirq))) {
#ifdef CONFIG_RCU_STATS
@@ -245,52 +558,82 @@ rcu_try_flip(void)
#endif /* #ifdef CONFIG_RCU_STATS */
		return;
	}
-	if (unlikely(flipctr != rcu_ctrlblk.completed)) {
-
-		/* Our work is done!  ;-) */
-
-#ifdef CONFIG_RCU_STATS
-		rcu_data.n_rcu_try_flip_e2++;
-#endif /* #ifdef CONFIG_RCU_STATS */
-		spin_unlock_irqrestore(&rcu_ctrlblk.fliplock, oldirq);
-		return;
-	}
-	flipctr &= 0x1;

	/*
-	 * Check for completion of all RCU read-side critical sections
-	 * that started prior to the previous flip.
+	 * Take the next transition(s) through the RCU grace-period
+	 * flip-counter state machine.
	 */

-#ifdef CONFIG_RCU_STATS
-	rcu_data.n_rcu_try_flip2++;
-#endif /* #ifdef CONFIG_RCU_STATS */
-	for_each_cpu(cpu) {
-		if (atomic_read(&per_cpu(rcu_flipctr, cpu)[!flipctr]) != 0) {
-#ifdef CONFIG_RCU_STATS
-			rcu_data.n_rcu_try_flip_e3++;
-#endif /* #ifdef CONFIG_RCU_STATS */
-			spin_unlock_irqrestore(&rcu_ctrlblk.fliplock,
oldirq);
-			return;
-		}
+	flipctr = rcu_ctrlblk.completed;
+	switch (rcu_try_flip_state) {
+	case rcu_try_flip_idle_state:
+		if (rcu_try_flip_idle(flipctr))
+			break;
+		rcu_try_flip_state = rcu_try_flip_in_gp_state;
+	case rcu_try_flip_in_gp_state:
+		if (rcu_try_flip_in_gp(flipctr))
+			break;
+		rcu_try_flip_state = rcu_try_flip_waitack_state;
+	case rcu_try_flip_waitack_state:
+		if (rcu_try_flip_waitack(flipctr))
+			break;
+		rcu_try_flip_state = rcu_try_flip_waitzero_state;
+	case rcu_try_flip_waitzero_state:
+		if (rcu_try_flip_waitzero(flipctr))
+			break;
+		rcu_try_flip_state = rcu_try_flip_waitmb_state;
+	case rcu_try_flip_waitmb_state:
+		if (rcu_try_flip_waitmb(flipctr))
+			break;
+		rcu_try_flip_state = rcu_try_flip_idle_state;
	}
+	spin_unlock_irqrestore(&rcu_ctrlblk.fliplock, oldirq);
+}

-	/* Do the flip. */
+/*
+ * Check to see if this CPU needs to report that it has seen the most
+ * recent counter flip, thereby declaring that all subsequent
+ * rcu_read_lock() invocations will respect this flip.
+ */

-	smp_mb();
-	rcu_ctrlblk.completed++;
+static void
+rcu_check_flipseen(int cpu)
+{
+	if (per_cpu(rcu_flip_flag, cpu) == rcu_flipped) {
+		smp_mb();  /* Subsequent counter acccesses must see new
value */
+		per_cpu(rcu_flip_flag, cpu) = rcu_flip_seen;
+		smp_mb();  /* probably be implied by interrupt, but... */
+	}
+}

-#ifdef CONFIG_RCU_STATS
-	rcu_data.n_rcu_try_flip3++;
-#endif /* #ifdef CONFIG_RCU_STATS */
-	spin_unlock_irqrestore(&rcu_ctrlblk.fliplock, oldirq);
+/*
+ * Check to see if this CPU needs to do a memory barrier in order to
+ * ensure that any prior RCU read-side critical sections have committed
+ * their counter manipulations and critical-section memory references
+ * before declaring the grace period to be completed.
+ */
+
+static void
+rcu_check_mb(int cpu)
+{
+	if (per_cpu(rcu_mb_flag, cpu) == rcu_mb_needed) {
+		smp_mb();
+		per_cpu(rcu_mb_flag, cpu) = rcu_mb_done;
+	}
}

+/*
+ * This function is called periodically on each CPU in hardware
interrupt
+ * context.
+ */
+
void
rcu_check_callbacks(int cpu, int user)
{
	unsigned long oldirq;

+	rcu_check_flipseen(cpu);
+	rcu_check_mb(cpu);
	if (rcu_ctrlblk.completed == rcu_data.completed) {
		rcu_try_flip();
		if (rcu_ctrlblk.completed == rcu_data.completed) {
@@ -359,10 +702,10 @@ call_rcu(struct rcu_head *head,
}

/*
- * Crude hack, reduces but does not eliminate possibility of failure.
- * Needs to wait for all CPUs to pass through a -voluntary- context
- * switch to eliminate possibility of failure.  (Maybe just crank
- * priority down...)
+ * Wait until all currently running preempt_disable() code segments
+ * (including hardware-irq-disable segments) complete.  Note that
+ * in -rt this does -not- necessarily result in all currently executing
+ * interrupt -handlers- having completed.
*/
void
synchronize_sched(void)
@@ -390,7 +733,7 @@ rcu_pending(int cpu)

void __init rcu_init(void)
{
-/*&&&&*/printk("WARNING: experimental RCU implementation.\n");
+/*&&&&*/printk("WARNING: experimental non-atomic RCU
implementation.\n");
	spin_lock_init(&rcu_data.lock);
	rcu_data.completed = 0;
	rcu_data.nextlist = NULL;
@@ -416,7 +759,8 @@ int rcu_read_proc_data(char *page)
	return sprintf(page,
		       "ggp=%ld lgp=%ld rcc=%ld\n"
		       "na=%ld nl=%ld wa=%ld wl=%ld da=%ld dl=%ld dr=%ld
		       di=%d\n"
-		       "rtf1=%d rtf2=%ld rtf3=%ld rtfe1=%d rtfe2=%ld
rtfe3=%ld\n",
+		       "1=%d e1=%d i1=%ld ie1=%ld g1=%ld a1=%ld ae1=%ld
a2=%ld\n"
+		       "z1=%ld ze1=%ld z2=%ld m1=%ld me1=%ld m2=%ld\n",

		       rcu_ctrlblk.completed,
		       rcu_data.completed,
@@ -431,12 +775,20 @@ int rcu_read_proc_data(char *page)
		       rcu_data.n_done_remove,
		       atomic_read(&rcu_data.n_done_invoked),

-		       atomic_read(&rcu_data.n_rcu_try_flip1),
-		       rcu_data.n_rcu_try_flip2,
-		       rcu_data.n_rcu_try_flip3,
+		       atomic_read(&rcu_data.n_rcu_try_flip_1),
		       atomic_read(&rcu_data.n_rcu_try_flip_e1),
-		       rcu_data.n_rcu_try_flip_e2,
-		       rcu_data.n_rcu_try_flip_e3);
+		       rcu_data.n_rcu_try_flip_i1,
+		       rcu_data.n_rcu_try_flip_ie1,
+		       rcu_data.n_rcu_try_flip_g1,
+		       rcu_data.n_rcu_try_flip_a1,
+		       rcu_data.n_rcu_try_flip_ae1,
+		       rcu_data.n_rcu_try_flip_a2,
+		       rcu_data.n_rcu_try_flip_z1,
+		       rcu_data.n_rcu_try_flip_ze1,
+		       rcu_data.n_rcu_try_flip_z2,
+		       rcu_data.n_rcu_try_flip_m1,
+		       rcu_data.n_rcu_try_flip_me1,
+		       rcu_data.n_rcu_try_flip_m2);
}

int rcu_read_proc_gp_data(char *page)
@@ -464,14 +816,23 @@ int rcu_read_proc_ctrs_data(char *page)
	int cpu;
	int f = rcu_data.completed & 0x1;

-	cnt += sprintf(&page[cnt], "CPU last cur\n");
+	cnt += sprintf(&page[cnt], "CPU last cur F M\n");
	for_each_cpu(cpu) {
-		cnt += sprintf(&page[cnt], "%3d %4d %3d\n",
+		cnt += sprintf(&page[cnt], "%3d %4d %3d %d %d\n",
			       cpu,
-			       atomic_read(&per_cpu(rcu_flipctr, cpu)[!f]),
-			       atomic_read(&per_cpu(rcu_flipctr, cpu)[f]));
-	}
-	cnt += sprintf(&page[cnt], "ggp = %ld\n", rcu_data.completed);
+			       per_cpu(rcu_flipctr, cpu)[!f],
+			       per_cpu(rcu_flipctr, cpu)[f],
+			       per_cpu(rcu_flip_flag, cpu),
+			       per_cpu(rcu_mb_flag, cpu));
+	}
+	cnt += sprintf(&page[cnt], "ggp = %ld, state = %d",
+		       rcu_data.completed, rcu_try_flip_state);
+	if ((0 <= rcu_try_flip_state) &&
+	    (rcu_try_flip_state <= sizeof(rcu_try_flip_state_names) /
+	    			   sizeof(rcu_try_flip_state_names[0])))
+		cnt += sprintf(&page[cnt], " (%s)",
+			       rcu_try_flip_state_names[rcu_try_flip_state]);
+	cnt += sprintf(&page[cnt], "\n");
	return (cnt);
}

-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to [email protected]
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/

---

• Follow-Ups:
  • Re: [RFC, PATCH -rt] NMI-safe mb- and atomic-free RT RCU
    • From: "Paul E. McKenney" <[email protected]>

• References:
  • [RFC, PATCH -rt] NMI-safe mb- and atomic-free RT RCU
    • From: "Paul E. McKenney" <[email protected]>
  • Re: [RFC, PATCH -rt] NMI-safe mb- and atomic-free RT RCU
    • From: Esben Nielsen <[email protected]>
  • Re: [RFC, PATCH -rt] NMI-safe mb- and atomic-free RT RCU
    • From: "Paul E. McKenney" <[email protected]>
  • Re: [RFC, PATCH -rt] NMI-safe mb- and atomic-free RT RCU
    • From: Esben Nielsen <[email protected]>
  • Re: [RFC, PATCH -rt] NMI-safe mb- and atomic-free RT RCU
    • From: "Paul E. McKenney" <[email protected]>

• Prev by Date: Re: [PATCH] [xfs] Add lock annotations to xfs_trans_update_ail and xfs_trans_delete_ail
• Next by Date: Re: suspend2 merge history [was Re: the " 'official' point of view" expressed by kernelnewbies.org regarding reiser4 inclusion]
• Previous by thread: Re: [RFC, PATCH -rt] NMI-safe mb- and atomic-free RT RCU
• Next by thread: Re: [RFC, PATCH -rt] NMI-safe mb- and atomic-free RT RCU
• Index(es):
  • Date
  • Thread

[Index of Archives]     [Kernel Newbies]     [Netfilter]     [Bugtraq]     [Photo]     [Stuff]     [Gimp]     [Yosemite News]     [MIPS Linux]     [ARM Linux]     [Linux Security]     [Linux RAID]     [Video 4 Linux]     [Linux for the blind]     [Linux Resources]