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.
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@arm.com>
Cc: Peter Zijlstra (Intel) <peterz@infradead.org>
Cc: Marco Elver <elver@google.com>
Cc: Frederic Weisbecker <frederic@kernel.org>
Cc: David Matlack <dmatlack@google.com>
Cc: Friedrich Weber <f.weber@proxmox.com>
Cc: Ankur Arora <ankur.a.arora@oracle.com>
Cc: Thomas Gleixner <tglx@linutronix.de>
Signed-off-by: Sean Christopherson <seanjc@google.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).
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
}
/*
--
2.44.0.278.ge034bb2e1d-googHi Sean,
On 2024-03-12 at 12:39:11 -0700, Sean Christopherson wrote:
> 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
It took me a while to wonder why PREEMPT_DYNAMIC selects PREEMPTION
in Kconfig, then I assume that you mean the static config is CONFIG_PREEMPTION,
but the live preemption model is "none" or "voluntary", which makes the
static check of CONFIG_PREEMPTION in spin_needbreak() and rwlock_needbreak()
invalid?
> 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.
>
> 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@arm.com>
> Cc: Peter Zijlstra (Intel) <peterz@infradead.org>
> Cc: Marco Elver <elver@google.com>
> Cc: Frederic Weisbecker <frederic@kernel.org>
> Cc: David Matlack <dmatlack@google.com>
> Cc: Friedrich Weber <f.weber@proxmox.com>
> Cc: Ankur Arora <ankur.a.arora@oracle.com>
> Cc: Thomas Gleixner <tglx@linutronix.de>
> Signed-off-by: Sean Christopherson <seanjc@google.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).
>
> 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())
The old version checks against static CONFIG_PREEMPTION, now we check
the live CONFIG_PREEMPTION and static CONFIG_PREEMPT_RT, just wonder
if the rt check is needed here?
thanks,
Chenyu
On Thu, Apr 25, 2024, Chen Yu wrote:
> Hi Sean,
>
> On 2024-03-12 at 12:39:11 -0700, Sean Christopherson wrote:
> > 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
>
> It took me a while to wonder why PREEMPT_DYNAMIC selects PREEMPTION
> in Kconfig, then I assume that you mean the static config is CONFIG_PREEMPTION,
> but the live preemption model is "none" or "voluntary", which makes the
> static check of CONFIG_PREEMPTION in spin_needbreak() and rwlock_needbreak()
> invalid?
Yep, exactly.
> > 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())
>
> The old version checks against static CONFIG_PREEMPTION, now we check
> the live CONFIG_PREEMPTION and static CONFIG_PREEMPT_RT, just wonder
> if the rt check is needed here?
It's required, as CONFIG_PREEMPT_RT=y doesn't imply CONFIG_PREEMPT, and
CONFIG_PREEMPT_RT=y is mutually exclusive with CONFIG_PREEMPT_DYNAMIC. I.e. a
CONFIG_PREEMPT_RT=y kernel will look yield:
CONFIG_PREEMPT_RT=y
CONFIG_PREEMPT_DYNAMIC=n
CONFIG_PREEMPT=n
which in turn generates:
static inline bool preempt_model_full(void)
{
return IS_ENABLED(CONFIG_PREEMPT);
}
and so just checking preempt_model_full() would incorrectly return false for
CONFIG_PREEMPT_RT=y.
On 2024-04-25 at 09:47:52 -0700, Sean Christopherson wrote:
> On Thu, Apr 25, 2024, Chen Yu wrote:
> > Hi Sean,
> >
> > On 2024-03-12 at 12:39:11 -0700, Sean Christopherson wrote:
> > > 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
> >
> > It took me a while to wonder why PREEMPT_DYNAMIC selects PREEMPTION
> > in Kconfig, then I assume that you mean the static config is CONFIG_PREEMPTION,
> > but the live preemption model is "none" or "voluntary", which makes the
> > static check of CONFIG_PREEMPTION in spin_needbreak() and rwlock_needbreak()
> > invalid?
>
> Yep, exactly.
>
> > > 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())
> >
> > The old version checks against static CONFIG_PREEMPTION, now we check
> > the live CONFIG_PREEMPTION and static CONFIG_PREEMPT_RT, just wonder
> > if the rt check is needed here?
>
> It's required, as CONFIG_PREEMPT_RT=y doesn't imply CONFIG_PREEMPT, and
> CONFIG_PREEMPT_RT=y is mutually exclusive with CONFIG_PREEMPT_DYNAMIC. I.e. a
> CONFIG_PREEMPT_RT=y kernel will look yield:
>
> CONFIG_PREEMPT_RT=y
> CONFIG_PREEMPT_DYNAMIC=n
> CONFIG_PREEMPT=n
>
> which in turn generates:
>
> static inline bool preempt_model_full(void)
> {
> return IS_ENABLED(CONFIG_PREEMPT);
> }
>
> and so just checking preempt_model_full() would incorrectly return false for
> CONFIG_PREEMPT_RT=y.
You are right, I missunderstood the definition of preempt_model_full(). For my
understanding of this patch:
Reviewed-by: Chen Yu <yu.c.chen@intel.com>
thanks,
Chenyu
Sean Christopherson <seanjc@google.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@arm.com>
> Cc: Peter Zijlstra (Intel) <peterz@infradead.org>
> Cc: Marco Elver <elver@google.com>
> Cc: Frederic Weisbecker <frederic@kernel.org>
> Cc: David Matlack <dmatlack@google.com>
> Cc: Friedrich Weber <f.weber@proxmox.com>
> Cc: Ankur Arora <ankur.a.arora@oracle.com>
> Cc: Thomas Gleixner <tglx@linutronix.de>
> Signed-off-by: Sean Christopherson <seanjc@google.com>
Reviewed-by: Ankur Arora <ankur.a.arora@oracle.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
> }
>
> /*
© 2016 - 2026 Red Hat, Inc.