linux-kernel - Re: [PATCH 2/5] sched: Add Lazy preemption model

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives
Hash Suite: Windows password security audit tool. GUI, reports in PDF.
[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <877cajdu8x.fsf@oracle.com>
Date: Mon, 07 Oct 2024 22:43:58 -0700
From: Ankur Arora <ankur.a.arora@...cle.com>
To: Peter Zijlstra <peterz@...radead.org>
Cc: bigeasy@...utronix.de, tglx@...utronix.de, mingo@...nel.org,
        linux-kernel@...r.kernel.org, juri.lelli@...hat.com,
        vincent.guittot@...aro.org, dietmar.eggemann@....com,
        rostedt@...dmis.org, bsegall@...gle.com, mgorman@...e.de,
        vschneid@...hat.com, ankur.a.arora@...cle.com, efault@....de
Subject: Re: [PATCH 2/5] sched: Add Lazy preemption model


Peter Zijlstra <peterz@...radead.org> writes:

> Change fair to use resched_curr_lazy(), which, when the lazy
> preemption model is selected, will set TIF_NEED_RESCHED_LAZY.
>
> This LAZY bit will be promoted to the full NEED_RESCHED bit on tick.
> As such, the average delay between setting LAZY and actually
> rescheduling will be TICK_NSEC/2.
>
> In short, Lazy preemption will delay preemption for fair class but
> will function as Full preemption for all the other classes, most
> notably the realtime (RR/FIFO/DEADLINE) classes.
>
> The goal is to bridge the performance gap with Voluntary, such that we
> might eventually remove that option entirely.
>
> Suggested-by: Thomas Gleixner <tglx@...utronix.de>
> Signed-off-by: Peter Zijlstra (Intel) <peterz@...radead.org>
> ---
>  include/linux/preempt.h |    8 ++++-
>  kernel/Kconfig.preempt  |   15 +++++++++
>  kernel/sched/core.c     |   76 ++++++++++++++++++++++++++++++++++++++++++++++--
>  kernel/sched/debug.c    |    5 +--
>  kernel/sched/fair.c     |    6 +--
>  kernel/sched/sched.h    |    1
>  6 files changed, 103 insertions(+), 8 deletions(-)
>
> --- a/include/linux/preempt.h
> +++ b/include/linux/preempt.h
> @@ -486,6 +486,7 @@ DEFINE_LOCK_GUARD_0(migrate, migrate_dis
>  extern bool preempt_model_none(void);
>  extern bool preempt_model_voluntary(void);
>  extern bool preempt_model_full(void);
> +extern bool preempt_model_lazy(void);
>
>  #else
>
> @@ -502,6 +503,11 @@ static inline bool preempt_model_full(vo
>  	return IS_ENABLED(CONFIG_PREEMPT);
>  }
>
> +static inline bool preempt_model_lazy(void)
> +{
> +	return IS_ENABLED(CONFIG_PREEMPT_LAZY);
> +}
> +
>  #endif
>
>  static inline bool preempt_model_rt(void)
> @@ -519,7 +525,7 @@ static inline bool preempt_model_rt(void
>   */
>  static inline bool preempt_model_preemptible(void)
>  {
> -	return preempt_model_full() || preempt_model_rt();
> +	return preempt_model_full() || preempt_model_lazy() || preempt_model_rt();
>  }

In addition to preempt_model_preemptible() we probably also need

  static inline bool preempt_model_minimize_latency(void)
  {
  	return preempt_model_full() || preempt_model_rt();
  }

for spin_needbreak()/rwlock_needbreak().

That would make the behaviour of spin_needbreak() under the lazy model
similar to none/voluntary.

>  #endif /* __LINUX_PREEMPT_H */
> --- a/kernel/Kconfig.preempt
> +++ b/kernel/Kconfig.preempt
> @@ -11,6 +11,9 @@ config PREEMPT_BUILD
>  	select PREEMPTION
>  	select UNINLINE_SPIN_UNLOCK if !ARCH_INLINE_SPIN_UNLOCK
>
> +config ARCH_HAS_PREEMPT_LAZY
> +	bool
> +
>  choice
>  	prompt "Preemption Model"
>  	default PREEMPT_NONE
> @@ -67,6 +70,18 @@ config PREEMPT
>  	  embedded system with latency requirements in the milliseconds
>  	  range.
>
> +config PREEMPT_LAZY
> +	bool "Scheduler controlled preemption model"
> +	depends on !ARCH_NO_PREEMPT
> +	depends on ARCH_HAS_PREEMPT_LAZY
> +	select PREEMPT_BUILD
> +	help
> +	  This option provides a scheduler driven preemption model that
> +	  is fundamentally similar to full preemption, but is less
> +	  eager to preempt SCHED_NORMAL tasks in an attempt to
> +	  reduce lock holder preemption and recover some of the performance
> +	  gains seen from using Voluntary preemption.
> +
>  config PREEMPT_RT
>  	bool "Fully Preemptible Kernel (Real-Time)"
>  	depends on EXPERT && ARCH_SUPPORTS_RT
> --- a/kernel/sched/core.c
> +++ b/kernel/sched/core.c
> @@ -1078,6 +1078,9 @@ static void __resched_curr(struct rq *rq
>
>  	lockdep_assert_rq_held(rq);
>
> +	if (is_idle_task(curr) && tif == TIF_NEED_RESCHED_LAZY)
> +		tif = TIF_NEED_RESCHED;
> +

Tasks with idle policy get handled at the usual user space boundary.
Maybe a comment reflecting that?

>  	if (cti->flags & ((1 << tif) | _TIF_NEED_RESCHED))
>  		return;
>
> @@ -1103,6 +1106,32 @@ void resched_curr(struct rq *rq)
>  	__resched_curr(rq, TIF_NEED_RESCHED);
>  }
>
> +#ifdef CONFIG_PREEMPT_DYNAMIC
> +static DEFINE_STATIC_KEY_FALSE(sk_dynamic_preempt_lazy);
> +static __always_inline bool dynamic_preempt_lazy(void)
> +{
> +	return static_branch_unlikely(&sk_dynamic_preempt_lazy);
> +}
> +#else
> +static __always_inline bool dynamic_preempt_lazy(void)
> +{
> +	return IS_ENABLED(PREEMPT_LAZY);
> +}
> +#endif
> +
> +static __always_inline int tif_need_resched_lazy(void)
> +{
> +	if (dynamic_preempt_lazy())
> +		return TIF_NEED_RESCHED_LAZY;
> +
> +	return TIF_NEED_RESCHED;
> +}

Nice. This simplifies things.

> +void resched_curr_lazy(struct rq *rq)
> +{
> +	__resched_curr(rq, tif_need_resched_lazy());
> +}
> +
>  void resched_cpu(int cpu)
>  {
>  	struct rq *rq = cpu_rq(cpu);
> @@ -5598,6 +5627,10 @@ void sched_tick(void)
>  	update_rq_clock(rq);
>  	hw_pressure = arch_scale_hw_pressure(cpu_of(rq));
>  	update_hw_load_avg(rq_clock_task(rq), rq, hw_pressure);
> +
> +	if (dynamic_preempt_lazy() && tif_test_bit(TIF_NEED_RESCHED_LAZY))
> +		resched_curr(rq);
> +

So this works for SCHED_NORMAL. But, does this do the right thing for
deadline etc other scheduling classes?


--
ankur