linux-kernel.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
From: "Paul E. McKenney" <paulmck@linux.vnet.ibm.com>
To: Lai Jiangshan <laijs@cn.fujitsu.com>
Cc: linux-kernel@vger.kernel.org, mingo@elte.hu, dipankar@in.ibm.com,
	akpm@linux-foundation.org, mathieu.desnoyers@polymtl.ca,
	josh@joshtriplett.org, niv@us.ibm.com, tglx@linutronix.de,
	peterz@infradead.org, rostedt@goodmis.org,
	Valdis.Kletnieks@vt.edu, dhowells@redhat.com,
	eric.dumazet@gmail.com, darren@dvhart.com, fweisbec@gmail.com,
	sbw@mit.edu, patches@linaro.org,
	"Paul E. McKenney" <paul.mckenney@linaro.org>,
	David Rientjes <rientjes@google.com>
Subject: Re: [PATCH tip/core/rcu 07/23] rcu: Provide OOM handler to motivate lazy RCU callbacks
Date: Wed, 5 Sep 2012 10:45:06 -0700	[thread overview]
Message-ID: <20120905174506.GL3308@linux.vnet.ibm.com> (raw)
In-Reply-To: <50447388.4080609@cn.fujitsu.com>

On Mon, Sep 03, 2012 at 05:08:24PM +0800, Lai Jiangshan wrote:
> On 08/31/2012 02:18 AM, Paul E. McKenney wrote:
> > From: "Paul E. McKenney" <paul.mckenney@linaro.org>
> > 
> > In kernels built with CONFIG_RCU_FAST_NO_HZ=y, CPUs can accumulate a
> > large number of lazy callbacks, which as the name implies will be slow
> > to be invoked.  This can be a problem on small-memory systems, where the
> > default 6-second sleep for CPUs having only lazy RCU callbacks could well
> > be fatal.  This commit therefore installs an OOM hander that ensures that
> > every CPU with non-lazy callbacks has at least one non-lazy callback,
> > in turn ensuring timely advancement for these callbacks.
> > 
> > Signed-off-by: Paul E. McKenney <paul.mckenney@linaro.org>
> > Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
> > Tested-by: Sasha Levin <levinsasha928@gmail.com>
> > ---
> >  kernel/rcutree.h        |    5 ++-
> >  kernel/rcutree_plugin.h |   80 +++++++++++++++++++++++++++++++++++++++++++++++
> >  2 files changed, 84 insertions(+), 1 deletions(-)
> > 
> > diff --git a/kernel/rcutree.h b/kernel/rcutree.h
> > index 117a150..effb273 100644
> > --- a/kernel/rcutree.h
> > +++ b/kernel/rcutree.h
> > @@ -315,8 +315,11 @@ struct rcu_data {
> >  	unsigned long n_rp_need_fqs;
> >  	unsigned long n_rp_need_nothing;
> >  
> > -	/* 6) _rcu_barrier() callback. */
> > +	/* 6) _rcu_barrier() and OOM callbacks. */
> >  	struct rcu_head barrier_head;
> > +#ifdef CONFIG_RCU_FAST_NO_HZ
> > +	struct rcu_head oom_head;
> > +#endif /* #ifdef CONFIG_RCU_FAST_NO_HZ */
> >  
> >  	int cpu;
> >  	struct rcu_state *rsp;
> > diff --git a/kernel/rcutree_plugin.h b/kernel/rcutree_plugin.h
> > index 7f3244c..bac8cc1 100644
> > --- a/kernel/rcutree_plugin.h
> > +++ b/kernel/rcutree_plugin.h
> > @@ -25,6 +25,7 @@
> >   */
> >  
> >  #include <linux/delay.h>
> > +#include <linux/oom.h>
> >  
> >  #define RCU_KTHREAD_PRIO 1
> >  
> > @@ -2112,6 +2113,85 @@ static void rcu_idle_count_callbacks_posted(void)
> >  	__this_cpu_add(rcu_dynticks.nonlazy_posted, 1);
> >  }
> >  
> > +/*
> > + * Data for flushing lazy RCU callbacks at OOM time.
> > + */
> > +static atomic_t oom_callback_count;
> > +static DECLARE_WAIT_QUEUE_HEAD(oom_callback_wq);
> > +
> > +/*
> > + * RCU OOM callback -- decrement the outstanding count and deliver the
> > + * wake-up if we are the last one.
> > + */
> > +static void rcu_oom_callback(struct rcu_head *rhp)
> > +{
> > +	if (atomic_dec_and_test(&oom_callback_count))
> > +		wake_up(&oom_callback_wq);
> > +}
> > +
> > +/*
> > + * Post an rcu_oom_notify callback on the current CPU if it has at
> > + * least one lazy callback.  This will unnecessarily post callbacks
> > + * to CPUs that already have a non-lazy callback at the end of their
> > + * callback list, but this is an infrequent operation, so accept some
> > + * extra overhead to keep things simple.
> > + */
> > +static void rcu_oom_notify_cpu(void *flavor)
> > +{
> > +	struct rcu_state *rsp = flavor;
> > +	struct rcu_data *rdp = __this_cpu_ptr(rsp->rda);
> > +
> > +	if (rdp->qlen_lazy != 0) {
> > +		atomic_inc(&oom_callback_count);
> > +		rsp->call(&rdp->oom_head, rcu_oom_callback);
> > +	}
> > +}
> > +
> > +/*
> > + * If low on memory, ensure that each CPU has a non-lazy callback.
> > + * This will wake up CPUs that have only lazy callbacks, in turn
> > + * ensuring that they free up the corresponding memory in a timely manner.
> > + */
> > +static int rcu_oom_notify(struct notifier_block *self,
> > +			  unsigned long notused, void *nfreed)
> > +{
> > +	int cpu;
> > +	struct rcu_state *rsp;
> > +
> > +	/* Wait for callbacks from earlier instance to complete. */
> > +	wait_event(oom_callback_wq, atomic_read(&oom_callback_count) == 0);
> > +
> > +	/*
> > +	 * Prevent premature wakeup: ensure that all increments happen
> > +	 * before there is a chance of the counter reaching zero.
> > +	 */
> > +	atomic_set(&oom_callback_count, 1);
> > +
> > +	get_online_cpus();
> > +	for_each_online_cpu(cpu)
> > +		for_each_rcu_flavor(rsp)
> > +			smp_call_function_single(cpu, rcu_oom_notify_cpu,
> > +						 rsp, 1);
> > +	put_online_cpus();
> > +
> > +	/* Unconditionally decrement: no need to wake ourselves up. */
> > +	atomic_dec(&oom_callback_count);
> > +
> > +	*(unsigned long *)nfreed = 1;
> 
> Hi, Paul
> 
> If you consider the above code has free some memory,
> you should use *(unsigned long *)nfreed = +1.
>                                           ^^
> 
> And your code disable OOM actually, because it transfer *nfreed to NON-ZERO
> unconditionally.

Hmmm...  That does indeed cause out_of_memory() to unconditionally
return, doesn't it?

So I should really just leave *nfreed alone, since I cannot be sure
whether or not anything will actually get freed.  I -could- count
callbacks, but they might well be allocated as fast as they are freed.

Good catch!!!

> I did not review the patch nor the whole series carefully.
> 
> And if it is possible, could you share the code with rcu_barrier()?

At the moment, it adds more code than it saves.

							Thanx, Paul

> Thanks,
> Lai
> 
> > +	return NOTIFY_OK;
> > +}
> > +
> > +static struct notifier_block rcu_oom_nb = {
> > +	.notifier_call = rcu_oom_notify
> > +};
> > +
> > +static int __init rcu_register_oom_notifier(void)
> > +{
> > +	register_oom_notifier(&rcu_oom_nb);
> > +	return 0;
> > +}
> > +early_initcall(rcu_register_oom_notifier);
> > +
> >  #endif /* #else #if !defined(CONFIG_RCU_FAST_NO_HZ) */
> >  
> >  #ifdef CONFIG_RCU_CPU_STALL_INFO
> 


  reply	other threads:[~2012-09-05 17:45 UTC|newest]

Thread overview: 85+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2012-08-30 18:18 [PATCH tip/core/rcu 0/23] Improvements to RT response on big systems and expedited functions Paul E. McKenney
2012-08-30 18:18 ` [PATCH tip/core/rcu 01/23] rcu: Move RCU grace-period initialization into a kthread Paul E. McKenney
2012-08-30 18:18   ` [PATCH tip/core/rcu 02/23] rcu: Allow RCU grace-period initialization to be preempted Paul E. McKenney
2012-09-02  1:09     ` Josh Triplett
2012-09-05  1:22       ` Paul E. McKenney
2012-08-30 18:18   ` [PATCH tip/core/rcu 03/23] rcu: Move RCU grace-period cleanup into kthread Paul E. McKenney
2012-09-02  1:22     ` Josh Triplett
2012-09-06 13:34     ` Peter Zijlstra
2012-09-06 17:29       ` Paul E. McKenney
2012-08-30 18:18   ` [PATCH tip/core/rcu 04/23] rcu: Allow RCU grace-period cleanup to be preempted Paul E. McKenney
2012-09-02  1:36     ` Josh Triplett
2012-08-30 18:18   ` [PATCH tip/core/rcu 05/23] rcu: Prevent offline CPUs from executing RCU core code Paul E. McKenney
2012-09-02  1:45     ` Josh Triplett
2012-08-30 18:18   ` [PATCH tip/core/rcu 06/23] rcu: Break up rcu_gp_kthread() into subfunctions Paul E. McKenney
2012-09-02  2:11     ` Josh Triplett
2012-09-06 13:39     ` Peter Zijlstra
2012-09-06 17:32       ` Paul E. McKenney
2012-09-06 18:49         ` Josh Triplett
2012-09-06 19:09           ` Peter Zijlstra
2012-09-06 20:30             ` Paul E. McKenney
2012-09-06 20:30           ` Paul E. McKenney
2012-08-30 18:18   ` [PATCH tip/core/rcu 07/23] rcu: Provide OOM handler to motivate lazy RCU callbacks Paul E. McKenney
2012-09-02  2:13     ` Josh Triplett
2012-09-03  9:08     ` Lai Jiangshan
2012-09-05 17:45       ` Paul E. McKenney [this message]
2012-09-06 13:46     ` Peter Zijlstra
2012-09-06 13:52       ` Steven Rostedt
2012-09-06 17:41         ` Paul E. McKenney
2012-09-06 17:46           ` Peter Zijlstra
2012-09-06 20:32             ` Paul E. McKenney
2012-08-30 18:18   ` [PATCH tip/core/rcu 08/23] rcu: Segregate rcu_state fields to improve cache locality Paul E. McKenney
2012-09-02  2:51     ` Josh Triplett
2012-08-30 18:18   ` [PATCH tip/core/rcu 09/23] rcu: Move quiescent-state forcing into kthread Paul E. McKenney
2012-08-30 18:18   ` [PATCH tip/core/rcu 10/23] rcu: Allow RCU quiescent-state forcing to be preempted Paul E. McKenney
2012-09-02  5:23     ` Josh Triplett
2012-08-30 18:18   ` [PATCH tip/core/rcu 11/23] rcu: Adjust debugfs tracing for kthread-based quiescent-state forcing Paul E. McKenney
2012-09-02  6:05     ` Josh Triplett
2012-08-30 18:18   ` [PATCH tip/core/rcu 12/23] rcu: Prevent force_quiescent_state() memory contention Paul E. McKenney
2012-09-02 10:47     ` Josh Triplett
2012-08-30 18:18   ` [PATCH tip/core/rcu 13/23] rcu: Control grace-period duration from sysfs Paul E. McKenney
2012-09-03  9:30     ` Josh Triplett
2012-09-03  9:31       ` Josh Triplett
2012-09-06 14:15     ` Peter Zijlstra
2012-09-06 17:53       ` Paul E. McKenney
2012-09-06 18:28         ` Peter Zijlstra
2012-09-06 20:37           ` Paul E. McKenney
2012-08-30 18:18   ` [PATCH tip/core/rcu 14/23] rcu: Remove now-unused rcu_state fields Paul E. McKenney
2012-09-03  9:31     ` Josh Triplett
2012-09-06 14:17     ` Peter Zijlstra
2012-09-06 18:02       ` Paul E. McKenney
2012-08-30 18:18   ` [PATCH tip/core/rcu 15/23] rcu: Make rcutree module parameters visible in sysfs Paul E. McKenney
2012-09-03  9:32     ` Josh Triplett
2012-08-30 18:18   ` [PATCH tip/core/rcu 16/23] rcu: Prevent initialization-time quiescent-state race Paul E. McKenney
2012-09-03  9:37     ` Josh Triplett
2012-09-05 18:19       ` Paul E. McKenney
2012-09-05 18:55         ` Josh Triplett
2012-09-05 19:49           ` Paul E. McKenney
2012-09-06 14:21         ` Peter Zijlstra
2012-09-06 16:18           ` Paul E. McKenney
2012-09-06 16:22             ` Peter Zijlstra
2012-08-30 18:18   ` [PATCH tip/core/rcu 17/23] rcu: Fix day-zero grace-period initialization/cleanup race Paul E. McKenney
2012-09-03  9:39     ` Josh Triplett
2012-09-06 14:24     ` Peter Zijlstra
2012-09-06 18:06       ` Paul E. McKenney
2012-08-30 18:18   ` [PATCH tip/core/rcu 18/23] rcu: Add random PROVE_RCU_DELAY to grace-period initialization Paul E. McKenney
2012-09-03  9:41     ` Josh Triplett
2012-09-06 14:27     ` Peter Zijlstra
2012-09-06 18:25       ` Paul E. McKenney
2012-08-30 18:18   ` [PATCH tip/core/rcu 19/23] rcu: Adjust for unconditional ->completed assignment Paul E. McKenney
2012-09-03  9:42     ` Josh Triplett
2012-08-30 18:18   ` [PATCH tip/core/rcu 20/23] rcu: Remove callback acceleration from grace-period initialization Paul E. McKenney
2012-09-03  9:42     ` Josh Triplett
2012-08-30 18:18   ` [PATCH tip/core/rcu 21/23] rcu: Eliminate signed overflow in synchronize_rcu_expedited() Paul E. McKenney
2012-09-03  9:43     ` Josh Triplett
2012-08-30 18:18   ` [PATCH tip/core/rcu 22/23] rcu: Reduce synchronize_rcu_expedited() latency Paul E. McKenney
2012-09-03  9:46     ` Josh Triplett
2012-08-30 18:18   ` [PATCH tip/core/rcu 23/23] rcu: Simplify quiescent-state detection Paul E. McKenney
2012-09-03  9:56     ` Josh Triplett
2012-09-06 14:36     ` Peter Zijlstra
2012-09-06 20:01       ` Paul E. McKenney
2012-09-06 21:18         ` Mathieu Desnoyers
2012-09-06 21:31           ` Paul E. McKenney
2012-09-02  1:04   ` [PATCH tip/core/rcu 01/23] rcu: Move RCU grace-period initialization into a kthread Josh Triplett
2012-09-06 13:32   ` Peter Zijlstra
2012-09-06 17:00     ` Paul E. McKenney

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=20120905174506.GL3308@linux.vnet.ibm.com \
    --to=paulmck@linux.vnet.ibm.com \
    --cc=Valdis.Kletnieks@vt.edu \
    --cc=akpm@linux-foundation.org \
    --cc=darren@dvhart.com \
    --cc=dhowells@redhat.com \
    --cc=dipankar@in.ibm.com \
    --cc=eric.dumazet@gmail.com \
    --cc=fweisbec@gmail.com \
    --cc=josh@joshtriplett.org \
    --cc=laijs@cn.fujitsu.com \
    --cc=linux-kernel@vger.kernel.org \
    --cc=mathieu.desnoyers@polymtl.ca \
    --cc=mingo@elte.hu \
    --cc=niv@us.ibm.com \
    --cc=patches@linaro.org \
    --cc=paul.mckenney@linaro.org \
    --cc=peterz@infradead.org \
    --cc=rientjes@google.com \
    --cc=rostedt@goodmis.org \
    --cc=sbw@mit.edu \
    --cc=tglx@linutronix.de \
    --subject='Re: [PATCH tip/core/rcu 07/23] rcu: Provide OOM handler to motivate lazy RCU callbacks' \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).