Message-ID: <87cyqedkyt.fsf@oracle.com>
Date: Wed, 24 Apr 2024 23:18:02 -0700
From: Ankur Arora <ankur.a.arora@...cle.com>
To: Sean Christopherson <seanjc@...gle.com>
Cc: Jonathan Corbet <corbet@....net>, Ingo Molnar <mingo@...hat.com>,
        Peter
 Zijlstra <peterz@...radead.org>,
        Juri Lelli <juri.lelli@...hat.com>,
        Vincent Guittot <vincent.guittot@...aro.org>,
        Will Deacon
 <will@...nel.org>, linux-doc@...r.kernel.org,
        linux-kernel@...r.kernel.org,
        Valentin Schneider
 <valentin.schneider@....com>,
        Marco Elver <elver@...gle.com>,
        Frederic
 Weisbecker <frederic@...nel.org>,
        David Matlack <dmatlack@...gle.com>,
        Friedrich Weber <f.weber@...xmox.com>,
        Ankur Arora
 <ankur.a.arora@...cle.com>,
        Thomas Gleixner <tglx@...utronix.de>
Subject: Re: [PATCH v2 2/2] sched/core: Drop spinlocks on contention iff
 kernel is preemptible


Sean Christopherson <seanjc@...gle.com> writes:

> Use preempt_model_preemptible() to detect a preemptible kernel when
> deciding whether or not to reschedule in order to drop a contended
> spinlock or rwlock.  Because PREEMPT_DYNAMIC selects PREEMPTION, kernels
> built with PREEMPT_DYNAMIC=y will yield contended locks even if the live
> preemption model is "none" or "voluntary".  In short, make kernels with
> dynamically selected models behave the same as kernels with statically
> selected models.

Agreed. This behaviour makes sense. Should also be useful for PREEMPT_AUTO.

The only thing that gives me pause is that now there is an extra
call+ret even when we don't yield the lock.

But maybe that could be addressed separately by converting
preempt_model_* to use a static key or similar.

> Somewhat counter-intuitively, NOT yielding a lock can provide better
> latency for the relevant tasks/processes.  E.g. KVM x86's mmu_lock, a
> rwlock, is often contended between an invalidation event (takes mmu_lock
> for write) and a vCPU servicing a guest page fault (takes mmu_lock for
> read).  For _some_ setups, letting the invalidation task complete even
> if there is mmu_lock contention provides lower latency for *all* tasks,
> i.e. the invalidation completes sooner *and* the vCPU services the guest
> page fault sooner.
>
> But even KVM's mmu_lock behavior isn't uniform, e.g. the "best" behavior
> can vary depending on the host VMM, the guest workload, the number of
> vCPUs, the number of pCPUs in the host, why there is lock contention, etc.
>
> In other words, simply deleting the CONFIG_PREEMPTION guard (or doing the
> opposite and removing contention yielding entirely) needs to come with a
> big pile of data proving that changing the status quo is a net positive.
>
> Opportunistically document this side effect of preempt=full, as yielding
> contended spinlocks can have significant, user-visible impact.
>
> Fixes: c597bfddc9e9 ("sched: Provide Kconfig support for default dynamic preempt mode")
> Link: https://lore.kernel.org/kvm/ef81ff36-64bb-4cfe-ae9b-e3acf47bff24@proxmox.com
> Cc: Valentin Schneider <valentin.schneider@....com>
> Cc: Peter Zijlstra (Intel) <peterz@...radead.org>
> Cc: Marco Elver <elver@...gle.com>
> Cc: Frederic Weisbecker <frederic@...nel.org>
> Cc: David Matlack <dmatlack@...gle.com>
> Cc: Friedrich Weber <f.weber@...xmox.com>
> Cc: Ankur Arora <ankur.a.arora@...cle.com>
> Cc: Thomas Gleixner <tglx@...utronix.de>
> Signed-off-by: Sean Christopherson <seanjc@...gle.com>

Reviewed-by: Ankur Arora <ankur.a.arora@...cle.com>

> ---
>  Documentation/admin-guide/kernel-parameters.txt |  4 +++-
>  include/linux/spinlock.h                        | 14 ++++++--------
>  2 files changed, 9 insertions(+), 9 deletions(-)
>
> diff --git a/Documentation/admin-guide/kernel-parameters.txt b/Documentation/admin-guide/kernel-parameters.txt
> index 825398d66c69..fdeddb066439 100644
> --- a/Documentation/admin-guide/kernel-parameters.txt
> +++ b/Documentation/admin-guide/kernel-parameters.txt
> @@ -4689,7 +4689,9 @@
>  			none - Limited to cond_resched() calls
>  			voluntary - Limited to cond_resched() and might_sleep() calls
>  			full - Any section that isn't explicitly preempt disabled
> -			       can be preempted anytime.
> +			       can be preempted anytime.  Tasks will also yield
> +			       contended spinlocks (if the critical section isn't
> +			       explicitly preempt disabled beyond the lock itself).

This seems to read a bit better:

+			       can be preempted anytime.  Tasks will also yield
+			       contended spinlocks (unless the critical section is
+			       explicitly preempt disabled beyond the lock itself).


Ankur

>  	print-fatal-signals=
>  			[KNL] debug: print fatal signals
> diff --git a/include/linux/spinlock.h b/include/linux/spinlock.h
> index 3fcd20de6ca8..63dd8cf3c3c2 100644
> --- a/include/linux/spinlock.h
> +++ b/include/linux/spinlock.h
> @@ -462,11 +462,10 @@ static __always_inline int spin_is_contended(spinlock_t *lock)
>   */
>  static inline int spin_needbreak(spinlock_t *lock)
>  {
> -#ifdef CONFIG_PREEMPTION
> +	if (!preempt_model_preemptible())
> +		return 0;
> +
>  	return spin_is_contended(lock);
> -#else
> -	return 0;
> -#endif
>  }
>
>  /*
> @@ -479,11 +478,10 @@ static inline int spin_needbreak(spinlock_t *lock)
>   */
>  static inline int rwlock_needbreak(rwlock_t *lock)
>  {
> -#ifdef CONFIG_PREEMPTION
> +	if (!preempt_model_preemptible())
> +		return 0;
> +
>  	return rwlock_is_contended(lock);
> -#else
> -	return 0;
> -#endif
>  }
>
>  /*

Powered by blists - more mailing lists