Re: [PATCH] rcu: Allow to eliminate softirq processing from rcutree

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

On Tue, Mar 19, 2019 at 08:26:13PM -0400, Joel Fernandes wrote:
> Adding the rcu@vger.kernel.org list as well, more comment below:
> 
> On Fri, Mar 15, 2019 at 12:11:30PM +0100, Sebastian Andrzej Siewior wrote:
> > From: "Paul E. McKenney" <paulmck@linux.vnet.ibm.com>
> > 
> > Running RCU out of softirq is a problem for some workloads that would
> > like to manage RCU core processing independently of other softirq work,
> > for example, setting kthread priority.
> > This commit therefore introduces the `rcunosoftirq' option which moves
> > the RCU core work from softirq to a per-CPU/per-flavor SCHED_OTHER
> > kthread named rcuc.
> > The SCHED_OTHER approach avoids the scalability problems that appeared
> > with the earlier attempt to move RCU core processing to from softirq to
> > kthreads.
> > That said, kernels built with RCU_BOOST=y will run the rcuc kthreads at
> > the RCU-boosting priority.
> > 
> > Reported-by: Thomas Gleixner <tglx@linutronix.de>
> > Tested-by: Mike Galbraith <bitbucket@online.de>
> > Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
> > [bigeasy: add rcunosoftirq option]
> > Signed-off-by: Sebastian Andrzej Siewior <bigeasy@linutronix.de>
> > ---
> >  kernel/rcu/tree.c        | 132 ++++++++++++++++++++++++++++++++---
> >  kernel/rcu/tree.h        |   4 +-
> >  kernel/rcu/tree_plugin.h | 145 +++++----------------------------------
> >  3 files changed, 141 insertions(+), 140 deletions(-)
> > 
> > diff --git a/kernel/rcu/tree.c b/kernel/rcu/tree.c
> > index 9180158756d2c..498dc5e9287d0 100644
> > --- a/kernel/rcu/tree.c
> > +++ b/kernel/rcu/tree.c
> > @@ -62,6 +62,12 @@
> >  #include <linux/suspend.h>
> >  #include <linux/ftrace.h>
> >  #include <linux/tick.h>
> > +#include <linux/gfp.h>
> > +#include <linux/oom.h>
> > +#include <linux/smpboot.h>
> > +#include <linux/jiffies.h>
> > +#include <linux/sched/isolation.h>
> > +#include "../time/tick-internal.h"
> >  
> >  #include "tree.h"
> >  #include "rcu.h"
> > @@ -2716,7 +2722,7 @@ EXPORT_SYMBOL_GPL(rcu_fwd_progress_check);
> >   * structures.  This may be called only from the CPU to whom the rdp
> >   * belongs.
> >   */
> > -static __latent_entropy void rcu_process_callbacks(struct softirq_action *unused)
> > +static __latent_entropy void rcu_process_callbacks(void)
> >  {
> >  	unsigned long flags;
> >  	struct rcu_data *rdp = raw_cpu_ptr(&rcu_data);
> > @@ -2758,6 +2764,13 @@ static __latent_entropy void rcu_process_callbacks(struct softirq_action *unused
> >  	trace_rcu_utilization(TPS("End RCU core"));
> >  }
> >  
> > +static void rcu_process_callbacks_si(struct softirq_action *h)
> > +{
> > +	rcu_process_callbacks();
> > +}
> > +
> > +static DEFINE_PER_CPU(struct task_struct *, rcu_cpu_kthread_task);
> > +
> >  /*
> >   * Schedule RCU callback invocation.  If the running implementation of RCU
> >   * does not support RCU priority boosting, just do a direct call, otherwise
> > @@ -2769,19 +2782,121 @@ static void invoke_rcu_callbacks(struct rcu_data *rdp)
> >  {
> >  	if (unlikely(!READ_ONCE(rcu_scheduler_fully_active)))
> >  		return;
> > -	if (likely(!rcu_state.boost)) {
> > -		rcu_do_batch(rdp);
> > -		return;
> > -	}
> > -	invoke_rcu_callbacks_kthread();
> > +	rcu_do_batch(rdp);
> 
> Looks like a nice change, but one question...
> 
> Consider the case where rcunosoftirq boot option is not passed.
> 
> Before, if RCU_BOOST=y, then callbacks would be invoked in rcuc threads if
> possible, by those threads being woken up from within the softirq context
> (in invoke_rcu_callbacks).
> 
> Now, if RCU_BOOST=y, then callbacks would only be invoked in softirq context
> and not in the threads at all. Because rcu_softirq_enabled = false, so the
> path executes:
>   rcu_read_unlock_special() ->
>         raise_softirq_irqsoff() ->
>                 rcu_process_callbacks_si() ->
>                         rcu_process_callbacks() ->
>                                 invoke_rcu_callbacks() ->
>                                         rcu_do_batch()
> 
> This seems like a behavioral change to me. This makes the callbacks always
> execute from the softirq context and not the threads when boosting is
> configured. IMO in the very least, such behavioral change should be
> documented in the change.
> 
> One way to fix this I think could be, if boosting is enabled, then set
> rcu_softirq_enabled to false by default so the callbacks are still executed
> in the rcuc threads.
> 
> Did I miss something? Sorry if I did, thanks!

This did pass light rcutorture testing without rcunosoftirq set, but it
is quite possible that it was just getting lucky.  Then again, there are
a few different versions of this patch in flight, as Sebastian notes in
his later email.  The one I used is on -rcu.

							Thanx, Paul

>  - Joel
> 
> 
> >  }
> >  
> > +static void rcu_wake_cond(struct task_struct *t, int status)
> > +{
> > +	/*
> > +	 * If the thread is yielding, only wake it when this
> > +	 * is invoked from idle
> > +	 */
> > +	if (t && (status != RCU_KTHREAD_YIELDING || is_idle_task(current)))
> > +		wake_up_process(t);
> > +}
> > +
> > +static bool rcu_softirq_enabled = true;
> > +
> > +static int __init rcunosoftirq_setup(char *str)
> > +{
> > +	rcu_softirq_enabled = false;
> > +	return 0;
> > +}
> > +__setup("rcunosoftirq", rcunosoftirq_setup);
> > +
> > +/*
> > + * Wake up this CPU's rcuc kthread to do RCU core processing.
> > + */
> >  static void invoke_rcu_core(void)
> >  {
> > -	if (cpu_online(smp_processor_id()))
> > +	unsigned long flags;
> > +	struct task_struct *t;
> > +
> > +	if (!cpu_online(smp_processor_id()))
> > +		return;
> > +	if (rcu_softirq_enabled) {
> >  		raise_softirq(RCU_SOFTIRQ);
> > +	} else {
> > +		local_irq_save(flags);
> > +		__this_cpu_write(rcu_cpu_has_work, 1);
> > +		t = __this_cpu_read(rcu_cpu_kthread_task);
> > +		if (t != NULL && current != t)
> > +			rcu_wake_cond(t, __this_cpu_read(rcu_cpu_kthread_status));
> > +		local_irq_restore(flags);
> > +	}
> >  }
> >  
> > +static void rcu_cpu_kthread_park(unsigned int cpu)
> > +{
> > +	per_cpu(rcu_cpu_kthread_status, cpu) = RCU_KTHREAD_OFFCPU;
> > +}
> > +
> > +static int rcu_cpu_kthread_should_run(unsigned int cpu)
> > +{
> > +	return __this_cpu_read(rcu_cpu_has_work);
> > +}
> > +
> > +/*
> > + * Per-CPU kernel thread that invokes RCU callbacks.  This replaces
> > + * the RCU softirq used in configurations of RCU that do not support RCU
> > + * priority boosting.
> > + */
> > +static void rcu_cpu_kthread(unsigned int cpu)
> > +{
> > +	unsigned int *statusp = this_cpu_ptr(&rcu_cpu_kthread_status);
> > +	char work, *workp = this_cpu_ptr(&rcu_cpu_has_work);
> > +	int spincnt;
> > +
> > +	for (spincnt = 0; spincnt < 10; spincnt++) {
> > +		trace_rcu_utilization(TPS("Start CPU kthread@rcu_wait"));
> > +		local_bh_disable();
> > +		*statusp = RCU_KTHREAD_RUNNING;
> > +		this_cpu_inc(rcu_cpu_kthread_loops);
> > +		local_irq_disable();
> > +		work = *workp;
> > +		*workp = 0;
> > +		local_irq_enable();
> > +		if (work)
> > +			rcu_process_callbacks();
> > +		local_bh_enable();
> > +		if (*workp == 0) {
> > +			trace_rcu_utilization(TPS("End CPU kthread@rcu_wait"));
> > +			*statusp = RCU_KTHREAD_WAITING;
> > +			return;
> > +		}
> > +	}
> > +	*statusp = RCU_KTHREAD_YIELDING;
> > +	trace_rcu_utilization(TPS("Start CPU kthread@rcu_yield"));
> > +	schedule_timeout_interruptible(2);
> > +	trace_rcu_utilization(TPS("End CPU kthread@rcu_yield"));
> > +	*statusp = RCU_KTHREAD_WAITING;
> > +}
> > +
> > +static struct smp_hotplug_thread rcu_cpu_thread_spec = {
> > +	.store			= &rcu_cpu_kthread_task,
> > +	.thread_should_run	= rcu_cpu_kthread_should_run,
> > +	.thread_fn		= rcu_cpu_kthread,
> > +	.thread_comm		= "rcuc/%u",
> > +	.setup			= rcu_cpu_kthread_setup,
> > +	.park			= rcu_cpu_kthread_park,
> > +};
> > +
> > +/*
> > + * Spawn per-CPU RCU core processing kthreads.
> > + */
> > +static int __init rcu_spawn_core_kthreads(void)
> > +{
> > +	int cpu;
> > +
> > +	for_each_possible_cpu(cpu)
> > +		per_cpu(rcu_cpu_has_work, cpu) = 0;
> > +	if (!IS_ENABLED(CONFIG_RCU_BOOST) && !rcu_softirq_enabled)
> > +		return 0;
> > +	WARN_ONCE(smpboot_register_percpu_thread(&rcu_cpu_thread_spec), "%s: Could not start rcub kthread, OOM is now expected behavior\n", __func__);
> > +	return 0;
> > +}
> > +early_initcall(rcu_spawn_core_kthreads);
> > +
> >  /*
> >   * Handle any core-RCU processing required by a call_rcu() invocation.
> >   */
> > @@ -3777,7 +3892,8 @@ void __init rcu_init(void)
> >  	rcu_init_one();
> >  	if (dump_tree)
> >  		rcu_dump_rcu_node_tree();
> > -	open_softirq(RCU_SOFTIRQ, rcu_process_callbacks);
> > +	if (rcu_softirq_enabled)
> > +		open_softirq(RCU_SOFTIRQ, rcu_process_callbacks_si);
> >  
> >  	/*
> >  	 * We don't need protection against CPU-hotplug here because
> > diff --git a/kernel/rcu/tree.h b/kernel/rcu/tree.h
> > index d90b02b53c0ec..fb8fc6ecc391b 100644
> > --- a/kernel/rcu/tree.h
> > +++ b/kernel/rcu/tree.h
> > @@ -402,12 +402,10 @@ static const char *tp_rcu_varname __used __tracepoint_string = rcu_name;
> >  
> >  int rcu_dynticks_snap(struct rcu_data *rdp);
> >  
> > -#ifdef CONFIG_RCU_BOOST
> >  DECLARE_PER_CPU(unsigned int, rcu_cpu_kthread_status);
> >  DECLARE_PER_CPU(int, rcu_cpu_kthread_cpu);
> >  DECLARE_PER_CPU(unsigned int, rcu_cpu_kthread_loops);
> >  DECLARE_PER_CPU(char, rcu_cpu_has_work);
> > -#endif /* #ifdef CONFIG_RCU_BOOST */
> >  
> >  /* Forward declarations for rcutree_plugin.h */
> >  static void rcu_bootup_announce(void);
> > @@ -425,8 +423,8 @@ void call_rcu(struct rcu_head *head, rcu_callback_t func);
> >  static void dump_blkd_tasks(struct rcu_node *rnp, int ncheck);
> >  static void rcu_initiate_boost(struct rcu_node *rnp, unsigned long flags);
> >  static void rcu_preempt_boost_start_gp(struct rcu_node *rnp);
> > -static void invoke_rcu_callbacks_kthread(void);
> >  static bool rcu_is_callbacks_kthread(void);
> > +static void rcu_cpu_kthread_setup(unsigned int cpu);
> >  static void __init rcu_spawn_boost_kthreads(void);
> >  static void rcu_prepare_kthreads(int cpu);
> >  static void rcu_cleanup_after_idle(void);
> > diff --git a/kernel/rcu/tree_plugin.h b/kernel/rcu/tree_plugin.h
> > index 1b3dd2fc0cd64..b440d6ef45d16 100644
> > --- a/kernel/rcu/tree_plugin.h
> > +++ b/kernel/rcu/tree_plugin.h
> > @@ -24,17 +24,6 @@
> >   *	   Paul E. McKenney <paulmck@linux.vnet.ibm.com>
> >   */
> >  
> > -#include <linux/delay.h>
> > -#include <linux/gfp.h>
> > -#include <linux/oom.h>
> > -#include <linux/sched/debug.h>
> > -#include <linux/smpboot.h>
> > -#include <linux/sched/isolation.h>
> > -#include <uapi/linux/sched/types.h>
> > -#include "../time/tick-internal.h"
> > -
> > -#ifdef CONFIG_RCU_BOOST
> > -
> >  #include "../locking/rtmutex_common.h"
> >  
> >  /*
> > @@ -45,19 +34,6 @@ DEFINE_PER_CPU(unsigned int, rcu_cpu_kthread_status);
> >  DEFINE_PER_CPU(unsigned int, rcu_cpu_kthread_loops);
> >  DEFINE_PER_CPU(char, rcu_cpu_has_work);
> >  
> > -#else /* #ifdef CONFIG_RCU_BOOST */
> > -
> > -/*
> > - * Some architectures do not define rt_mutexes, but if !CONFIG_RCU_BOOST,
> > - * all uses are in dead code.  Provide a definition to keep the compiler
> > - * happy, but add WARN_ON_ONCE() to complain if used in the wrong place.
> > - * This probably needs to be excluded from -rt builds.
> > - */
> > -#define rt_mutex_owner(a) ({ WARN_ON_ONCE(1); NULL; })
> > -#define rt_mutex_futex_unlock(x) WARN_ON_ONCE(1)
> > -
> > -#endif /* #else #ifdef CONFIG_RCU_BOOST */
> > -
> >  #ifdef CONFIG_RCU_NOCB_CPU
> >  static cpumask_var_t rcu_nocb_mask; /* CPUs to have callbacks offloaded. */
> >  static bool __read_mostly rcu_nocb_poll;    /* Offload kthread are to poll. */
> > @@ -652,7 +628,10 @@ static void rcu_read_unlock_special(struct task_struct *t)
> >  		/* Need to defer quiescent state until everything is enabled. */
> >  		if (irqs_were_disabled) {
> >  			/* Enabling irqs does not reschedule, so... */
> > -			raise_softirq_irqoff(RCU_SOFTIRQ);
> > +			if (rcu_softirq_enabled)
> > +				raise_softirq_irqoff(RCU_SOFTIRQ);
> > +			else
> > +				invoke_rcu_core();
> >  		} else {
> >  			/* Enabling BH or preempt does reschedule, so... */
> >  			set_tsk_need_resched(current);
> > @@ -1150,18 +1129,21 @@ dump_blkd_tasks(struct rcu_node *rnp, int ncheck)
> >  
> >  #endif /* #else #ifdef CONFIG_PREEMPT_RCU */
> >  
> > -#ifdef CONFIG_RCU_BOOST
> > -
> > -static void rcu_wake_cond(struct task_struct *t, int status)
> > +/*
> > + * If boosting, set rcuc kthreads to realtime priority.
> > + */
> > +static void rcu_cpu_kthread_setup(unsigned int cpu)
> >  {
> > -	/*
> > -	 * If the thread is yielding, only wake it when this
> > -	 * is invoked from idle
> > -	 */
> > -	if (status != RCU_KTHREAD_YIELDING || is_idle_task(current))
> > -		wake_up_process(t);
> > +#ifdef CONFIG_RCU_BOOST
> > +	struct sched_param sp;
> > +
> > +	sp.sched_priority = kthread_prio;
> > +	sched_setscheduler_nocheck(current, SCHED_FIFO, &sp);
> > +#endif /* #ifdef CONFIG_RCU_BOOST */
> >  }
> >  
> > +#ifdef CONFIG_RCU_BOOST
> > +
> >  /*
> >   * Carry out RCU priority boosting on the task indicated by ->exp_tasks
> >   * or ->boost_tasks, advancing the pointer to the next task in the
> > @@ -1299,23 +1281,6 @@ static void rcu_initiate_boost(struct rcu_node *rnp, unsigned long flags)
> >  	}
> >  }
> >  
> > -/*
> > - * Wake up the per-CPU kthread to invoke RCU callbacks.
> > - */
> > -static void invoke_rcu_callbacks_kthread(void)
> > -{
> > -	unsigned long flags;
> > -
> > -	local_irq_save(flags);
> > -	__this_cpu_write(rcu_cpu_has_work, 1);
> > -	if (__this_cpu_read(rcu_cpu_kthread_task) != NULL &&
> > -	    current != __this_cpu_read(rcu_cpu_kthread_task)) {
> > -		rcu_wake_cond(__this_cpu_read(rcu_cpu_kthread_task),
> > -			      __this_cpu_read(rcu_cpu_kthread_status));
> > -	}
> > -	local_irq_restore(flags);
> > -}
> > -
> >  /*
> >   * Is the current CPU running the RCU-callbacks kthread?
> >   * Caller must have preemption disabled.
> > @@ -1369,65 +1334,6 @@ static int rcu_spawn_one_boost_kthread(struct rcu_node *rnp)
> >  	return 0;
> >  }
> >  
> > -static void rcu_kthread_do_work(void)
> > -{
> > -	rcu_do_batch(this_cpu_ptr(&rcu_data));
> > -}
> > -
> > -static void rcu_cpu_kthread_setup(unsigned int cpu)
> > -{
> > -	struct sched_param sp;
> > -
> > -	sp.sched_priority = kthread_prio;
> > -	sched_setscheduler_nocheck(current, SCHED_FIFO, &sp);
> > -}
> > -
> > -static void rcu_cpu_kthread_park(unsigned int cpu)
> > -{
> > -	per_cpu(rcu_cpu_kthread_status, cpu) = RCU_KTHREAD_OFFCPU;
> > -}
> > -
> > -static int rcu_cpu_kthread_should_run(unsigned int cpu)
> > -{
> > -	return __this_cpu_read(rcu_cpu_has_work);
> > -}
> > -
> > -/*
> > - * Per-CPU kernel thread that invokes RCU callbacks.  This replaces
> > - * the RCU softirq used in configurations of RCU that do not support RCU
> > - * priority boosting.
> > - */
> > -static void rcu_cpu_kthread(unsigned int cpu)
> > -{
> > -	unsigned int *statusp = this_cpu_ptr(&rcu_cpu_kthread_status);
> > -	char work, *workp = this_cpu_ptr(&rcu_cpu_has_work);
> > -	int spincnt;
> > -
> > -	for (spincnt = 0; spincnt < 10; spincnt++) {
> > -		trace_rcu_utilization(TPS("Start CPU kthread@rcu_wait"));
> > -		local_bh_disable();
> > -		*statusp = RCU_KTHREAD_RUNNING;
> > -		this_cpu_inc(rcu_cpu_kthread_loops);
> > -		local_irq_disable();
> > -		work = *workp;
> > -		*workp = 0;
> > -		local_irq_enable();
> > -		if (work)
> > -			rcu_kthread_do_work();
> > -		local_bh_enable();
> > -		if (*workp == 0) {
> > -			trace_rcu_utilization(TPS("End CPU kthread@rcu_wait"));
> > -			*statusp = RCU_KTHREAD_WAITING;
> > -			return;
> > -		}
> > -	}
> > -	*statusp = RCU_KTHREAD_YIELDING;
> > -	trace_rcu_utilization(TPS("Start CPU kthread@rcu_yield"));
> > -	schedule_timeout_interruptible(2);
> > -	trace_rcu_utilization(TPS("End CPU kthread@rcu_yield"));
> > -	*statusp = RCU_KTHREAD_WAITING;
> > -}
> > -
> >  /*
> >   * Set the per-rcu_node kthread's affinity to cover all CPUs that are
> >   * served by the rcu_node in question.  The CPU hotplug lock is still
> > @@ -1458,27 +1364,13 @@ static void rcu_boost_kthread_setaffinity(struct rcu_node *rnp, int outgoingcpu)
> >  	free_cpumask_var(cm);
> >  }
> >  
> > -static struct smp_hotplug_thread rcu_cpu_thread_spec = {
> > -	.store			= &rcu_cpu_kthread_task,
> > -	.thread_should_run	= rcu_cpu_kthread_should_run,
> > -	.thread_fn		= rcu_cpu_kthread,
> > -	.thread_comm		= "rcuc/%u",
> > -	.setup			= rcu_cpu_kthread_setup,
> > -	.park			= rcu_cpu_kthread_park,
> > -};
> > -
> >  /*
> >   * Spawn boost kthreads -- called as soon as the scheduler is running.
> >   */
> >  static void __init rcu_spawn_boost_kthreads(void)
> >  {
> >  	struct rcu_node *rnp;
> > -	int cpu;
> >  
> > -	for_each_possible_cpu(cpu)
> > -		per_cpu(rcu_cpu_has_work, cpu) = 0;
> > -	if (WARN_ONCE(smpboot_register_percpu_thread(&rcu_cpu_thread_spec), "%s: Could not start rcub kthread, OOM is now expected behavior\n", __func__))
> > -		return;
> >  	rcu_for_each_leaf_node(rnp)
> >  		(void)rcu_spawn_one_boost_kthread(rnp);
> >  }
> > @@ -1501,11 +1393,6 @@ static void rcu_initiate_boost(struct rcu_node *rnp, unsigned long flags)
> >  	raw_spin_unlock_irqrestore_rcu_node(rnp, flags);
> >  }
> >  
> > -static void invoke_rcu_callbacks_kthread(void)
> > -{
> > -	WARN_ON_ONCE(1);
> > -}
> > -
> >  static bool rcu_is_callbacks_kthread(void)
> >  {
> >  	return false;
> > -- 
> > 2.20.1
> > 
>