Re: mm,numad,rcu: hang on OOM

["Paul E. McKenney" <paulmck@linux.vnet.ibm.com>]

On Tue, Jul 17, 2012 at 10:15:46AM +0200, Sasha Levin wrote:
> Hi Paul,
> 
> I've been running with your patch below for a while now, and haven't encountered the issue again.

Thank you for the testing, Sasha!  I have the full OOM patch queued
for 3.7.

							Thanx, Paul

> On 07/01/2012 03:15 AM, Paul E. McKenney wrote:
> > On Fri, Jun 29, 2012 at 06:44:41PM +0200, Sasha Levin wrote:
> >> Hi all,
> >>
> >> While fuzzing using trinity on a KVM tools guest with todays linux-next, I've hit the following lockup:
> >>
> >> [  362.261729] INFO: task numad/2:27 blocked for more than 120 seconds.
> >> [  362.263974] "echo 0 > /proc/sys/kernel/hung_task_timeout_secs" disables this message.
> >> [  362.271684] numad/2         D 0000000000000001  5672    27      2 0x00000000
> >> [  362.280052]  ffff8800294c7c58 0000000000000046 ffff8800294c7c08 ffffffff81163dba
> >> [  362.294477]  ffff8800294c6000 ffff8800294c6010 ffff8800294c7fd8 ffff8800294c6000
> >> [  362.306631]  ffff8800294c6010 ffff8800294c7fd8 ffff88000d5c3000 ffff8800294c8000
> >> [  362.315395] Call Trace:
> >> [  362.318556]  [<ffffffff81163dba>] ? __lock_release+0x1ba/0x1d0
> >> [  362.325411]  [<ffffffff8372ab75>] schedule+0x55/0x60
> >> [  362.328844]  [<ffffffff8372b965>] rwsem_down_failed_common+0xf5/0x130
> >> [  362.332501]  [<ffffffff8115d38e>] ? put_lock_stats+0xe/0x40
> >> [  362.334496]  [<ffffffff81160135>] ? __lock_contended+0x1f5/0x230
> >> [  362.336723]  [<ffffffff8372b9d5>] rwsem_down_read_failed+0x15/0x17
> >> [  362.339297]  [<ffffffff81985e34>] call_rwsem_down_read_failed+0x14/0x30
> >> [  362.341768]  [<ffffffff83729a29>] ? down_read+0x79/0xa0
> >> [  362.343669]  [<ffffffff8122d262>] ? lazy_migrate_process+0x22/0x60
> >> [  362.345616]  [<ffffffff8122d262>] lazy_migrate_process+0x22/0x60
> >> [  362.347464]  [<ffffffff811453c0>] process_mem_migrate+0x10/0x20
> >> [  362.349340]  [<ffffffff81145090>] move_processes+0x190/0x230
> >> [  362.351398]  [<ffffffff81145b7a>] numad_thread+0x7a/0x120
> >> [  362.353245]  [<ffffffff81145b00>] ? find_busiest_node+0x310/0x310
> >> [  362.355396]  [<ffffffff81119e82>] kthread+0xb2/0xc0
> >> [  362.356996]  [<ffffffff8372ea34>] kernel_thread_helper+0x4/0x10
> >> [  362.359253]  [<ffffffff8372ccb4>] ? retint_restore_args+0x13/0x13
> >> [  362.361168]  [<ffffffff81119dd0>] ? __init_kthread_worker+0x70/0x70
> >> [  362.363277]  [<ffffffff8372ea30>] ? gs_change+0x13/0x13
> >>
> >> I've hit sysrq-t to see what might be the cause, and it appears that an OOM was in progress, and was stuck on RCU:
> >>
> >> [  578.086230] trinity-child69 D ffff8800277a54c8  3968  6658   6580 0x00000000
> >> [  578.086230]  ffff880022c5f518 0000000000000046 ffff880022c5f4c8 ffff88001b9d6e00
> >> [  578.086230]  ffff880022c5e000 ffff880022c5e010 ffff880022c5ffd8 ffff880022c5e000
> >> [  578.086230]  ffff880022c5e010 ffff880022c5ffd8 ffff880023c08000 ffff880022c33000
> >> [  578.086230] Call Trace:
> >> [  578.086230]  [<ffffffff8372ab75>] schedule+0x55/0x60
> >> [  578.086230]  [<ffffffff837285c8>] schedule_timeout+0x38/0x2c0
> >> [  578.086230]  [<ffffffff81161d16>] ? mark_held_locks+0xf6/0x120
> >> [  578.086230]  [<ffffffff81163dba>] ? __lock_release+0x1ba/0x1d0
> >> [  578.086230]  [<ffffffff8372c67b>] ? _raw_spin_unlock_irq+0x2b/0x80
> >> [  578.086230]  [<ffffffff8372a06f>] wait_for_common+0xff/0x170
> >> [  578.086230]  [<ffffffff81132c10>] ? try_to_wake_up+0x290/0x290
> >> [  578.086230]  [<ffffffff8372a188>] wait_for_completion+0x18/0x20
> >> [  578.086230]  [<ffffffff811a5de7>] _rcu_barrier+0x4a7/0x4e0
> > 
> > Hmmm...  Perhaps a blocking operation is not appropriate here.  I have
> > substituted a nonblocking approach, which is at -rcu (thus soon -next)
> > at 1ee4c09d (Provide OOM handler to motivate lazy RCU callbacks).
> > Patch below.
> > 
> > 							Thanx, Paul
> > 
> > ------------------------------------------------------------------------
> > 
> > rcu: Provide OOM handler to motivate lazy RCU callbacks
> > 
> > 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>
> > 
> > diff --git a/kernel/rcutree.h b/kernel/rcutree.h
> > index 4b47fbe..dab279f 100644
> > --- a/kernel/rcutree.h
> > +++ b/kernel/rcutree.h
> > @@ -314,8 +314,11 @@ 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 81e53eb..1908847 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
> >  
> > @@ -2128,6 +2129,90 @@ 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;
> > +
> > +	/* 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) {
> > +#ifdef CONFIG_PREEMPT_RCU
> > +		smp_call_function_single(cpu, rcu_oom_notify_cpu,
> > +					 &rcu_preempt_state, 1);
> > +#endif /* #ifdef CONFIG_PREEMPT_RCU */
> > +		smp_call_function_single(cpu, rcu_oom_notify_cpu,
> > +					 &rcu_bh_state, 1);
> > +		smp_call_function_single(cpu, rcu_oom_notify_cpu,
> > +					 &rcu_sched_state, 1);
> > +	}
> > +	put_online_cpus();
> > +
> > +	/* Unconditionally decrement: no need to wake ourselves up. */
> > +	atomic_dec(&oom_callback_count);
> > +
> > +	*(unsigned long *)nfreed = 1;
> > +	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
> > 
> > --
> > To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
> > the body of a message to majordomo@vger.kernel.org
> > More majordomo info at  http://vger.kernel.org/majordomo-info.html
> > Please read the FAQ at  http://www.tux.org/lkml/
> > 
> 
> 

--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@kvack.org.  For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"dont@kvack.org"> email@kvack.org </a>