From aa65ff6b18e0366db1790609956a4ac7308c5668 Mon Sep 17 00:00:00 2001 From: Nicholas Piggin Date: Fri, 24 Jul 2020 23:14:20 +1000 Subject: powerpc/64s: Implement queued spinlocks and rwlocks These have shown significantly improved performance and fairness when spinlock contention is moderate to high on very large systems. With this series including subsequent patches, on a 16 socket 1536 thread POWER9, a stress test such as same-file open/close from all CPUs gets big speedups, 11620op/s aggregate with simple spinlocks vs 384158op/s (33x faster), where the difference in throughput between the fastest and slowest thread goes from 7x to 1.4x. Thanks to the fast path being identical in terms of atomics and barriers (after a subsequent optimisation patch), single threaded performance is not changed (no measurable difference). On smaller systems, performance and fairness seems to be generally improved. Using dbench on tmpfs as a test (that starts to run into kernel spinlock contention), a 2-socket OpenPOWER POWER9 system was tested with bare metal and KVM guest configurations. Results can be found here: https://github.com/linuxppc/issues/issues/305#issuecomment-663487453 Observations are: - Queued spinlocks are equal when contention is insignificant, as expected and as measured with microbenchmarks. - When there is contention, on bare metal queued spinlocks have better throughput and max latency at all points. - When virtualised, queued spinlocks are slightly worse approaching peak throughput, but significantly better throughput and max latency at all points beyond peak, until queued spinlock maximum latency rises when clients are 2x vCPUs. The regressions haven't been analysed very well yet, there are a lot of things that can be tuned, particularly the paravirtualised locking, but the numbers already look like a good net win even on relatively small systems. Signed-off-by: Nicholas Piggin Acked-by: Peter Zijlstra (Intel) Acked-by: Waiman Long Signed-off-by: Michael Ellerman Link: https://lore.kernel.org/r/20200724131423.1362108-4-npiggin@gmail.com --- include/asm-generic/qspinlock.h | 2 ++ 1 file changed, 2 insertions(+) (limited to 'include/asm-generic/qspinlock.h') diff --git a/include/asm-generic/qspinlock.h b/include/asm-generic/qspinlock.h index fde943d180e0..fb0a814d4395 100644 --- a/include/asm-generic/qspinlock.h +++ b/include/asm-generic/qspinlock.h @@ -12,6 +12,7 @@ #include +#ifndef queued_spin_is_locked /** * queued_spin_is_locked - is the spinlock locked? * @lock: Pointer to queued spinlock structure @@ -25,6 +26,7 @@ static __always_inline int queued_spin_is_locked(struct qspinlock *lock) */ return atomic_read(&lock->val); } +#endif /** * queued_spin_value_unlocked - is the spinlock structure unlocked? -- cgit v1.2.3 From 20c0e8269e9d515e677670902c7e1cc0209d6ad9 Mon Sep 17 00:00:00 2001 From: Nicholas Piggin Date: Fri, 24 Jul 2020 23:14:21 +1000 Subject: powerpc/pseries: Implement paravirt qspinlocks for SPLPAR This implements the generic paravirt qspinlocks using H_PROD and H_CONFER to kick and wait. This uses an un-directed yield to any CPU rather than the directed yield to a pre-empted lock holder that paravirtualised simple spinlocks use, that requires no kick hcall. This is something that could be investigated and improved in future. Performance results can be found in the commit which added queued spinlocks. Signed-off-by: Nicholas Piggin Acked-by: Peter Zijlstra (Intel) Acked-by: Waiman Long Signed-off-by: Michael Ellerman Link: https://lore.kernel.org/r/20200724131423.1362108-5-npiggin@gmail.com --- arch/powerpc/include/asm/paravirt.h | 28 ++++++++++++ arch/powerpc/include/asm/qspinlock.h | 66 +++++++++++++++++++++++++++ arch/powerpc/include/asm/qspinlock_paravirt.h | 7 +++ arch/powerpc/include/asm/spinlock.h | 4 ++ arch/powerpc/platforms/pseries/Kconfig | 9 +++- arch/powerpc/platforms/pseries/setup.c | 4 +- include/asm-generic/qspinlock.h | 2 + 7 files changed, 118 insertions(+), 2 deletions(-) create mode 100644 arch/powerpc/include/asm/qspinlock_paravirt.h (limited to 'include/asm-generic/qspinlock.h') diff --git a/arch/powerpc/include/asm/paravirt.h b/arch/powerpc/include/asm/paravirt.h index 339e8533464b..9362c94fe3aa 100644 --- a/arch/powerpc/include/asm/paravirt.h +++ b/arch/powerpc/include/asm/paravirt.h @@ -28,6 +28,16 @@ static inline void yield_to_preempted(int cpu, u32 yield_count) { plpar_hcall_norets(H_CONFER, get_hard_smp_processor_id(cpu), yield_count); } + +static inline void prod_cpu(int cpu) +{ + plpar_hcall_norets(H_PROD, get_hard_smp_processor_id(cpu)); +} + +static inline void yield_to_any(void) +{ + plpar_hcall_norets(H_CONFER, -1, 0); +} #else static inline bool is_shared_processor(void) { @@ -44,6 +54,19 @@ static inline void yield_to_preempted(int cpu, u32 yield_count) { ___bad_yield_to_preempted(); /* This would be a bug */ } + +extern void ___bad_yield_to_any(void); +static inline void yield_to_any(void) +{ + ___bad_yield_to_any(); /* This would be a bug */ +} + +extern void ___bad_prod_cpu(void); +static inline void prod_cpu(int cpu) +{ + ___bad_prod_cpu(); /* This would be a bug */ +} + #endif #define vcpu_is_preempted vcpu_is_preempted @@ -56,4 +79,9 @@ static inline bool vcpu_is_preempted(int cpu) return false; } +static inline bool pv_is_native_spin_unlock(void) +{ + return !is_shared_processor(); +} + #endif /* _ASM_POWERPC_PARAVIRT_H */ diff --git a/arch/powerpc/include/asm/qspinlock.h b/arch/powerpc/include/asm/qspinlock.h index c49e33e24edd..f5066f00a08c 100644 --- a/arch/powerpc/include/asm/qspinlock.h +++ b/arch/powerpc/include/asm/qspinlock.h @@ -3,9 +3,47 @@ #define _ASM_POWERPC_QSPINLOCK_H #include +#include #define _Q_PENDING_LOOPS (1 << 9) /* not tuned */ +#ifdef CONFIG_PARAVIRT_SPINLOCKS +extern void native_queued_spin_lock_slowpath(struct qspinlock *lock, u32 val); +extern void __pv_queued_spin_lock_slowpath(struct qspinlock *lock, u32 val); +extern void __pv_queued_spin_unlock(struct qspinlock *lock); + +static __always_inline void queued_spin_lock_slowpath(struct qspinlock *lock, u32 val) +{ + if (!is_shared_processor()) + native_queued_spin_lock_slowpath(lock, val); + else + __pv_queued_spin_lock_slowpath(lock, val); +} + +#define queued_spin_unlock queued_spin_unlock +static inline void queued_spin_unlock(struct qspinlock *lock) +{ + if (!is_shared_processor()) + smp_store_release(&lock->locked, 0); + else + __pv_queued_spin_unlock(lock); +} + +#else +extern void queued_spin_lock_slowpath(struct qspinlock *lock, u32 val); +#endif + +static __always_inline void queued_spin_lock(struct qspinlock *lock) +{ + u32 val = 0; + + if (likely(atomic_try_cmpxchg_acquire(&lock->val, &val, _Q_LOCKED_VAL))) + return; + + queued_spin_lock_slowpath(lock, val); +} +#define queued_spin_lock queued_spin_lock + #define smp_mb__after_spinlock() smp_mb() static __always_inline int queued_spin_is_locked(struct qspinlock *lock) @@ -20,6 +58,34 @@ static __always_inline int queued_spin_is_locked(struct qspinlock *lock) } #define queued_spin_is_locked queued_spin_is_locked +#ifdef CONFIG_PARAVIRT_SPINLOCKS +#define SPIN_THRESHOLD (1<<15) /* not tuned */ + +static __always_inline void pv_wait(u8 *ptr, u8 val) +{ + if (*ptr != val) + return; + yield_to_any(); + /* + * We could pass in a CPU here if waiting in the queue and yield to + * the previous CPU in the queue. + */ +} + +static __always_inline void pv_kick(int cpu) +{ + prod_cpu(cpu); +} + +extern void __pv_init_lock_hash(void); + +static inline void pv_spinlocks_init(void) +{ + __pv_init_lock_hash(); +} + +#endif + #include #endif /* _ASM_POWERPC_QSPINLOCK_H */ diff --git a/arch/powerpc/include/asm/qspinlock_paravirt.h b/arch/powerpc/include/asm/qspinlock_paravirt.h new file mode 100644 index 000000000000..6b60e7736a47 --- /dev/null +++ b/arch/powerpc/include/asm/qspinlock_paravirt.h @@ -0,0 +1,7 @@ +/* SPDX-License-Identifier: GPL-2.0-or-later */ +#ifndef _ASM_POWERPC_QSPINLOCK_PARAVIRT_H +#define _ASM_POWERPC_QSPINLOCK_PARAVIRT_H + +EXPORT_SYMBOL(__pv_queued_spin_unlock); + +#endif /* _ASM_POWERPC_QSPINLOCK_PARAVIRT_H */ diff --git a/arch/powerpc/include/asm/spinlock.h b/arch/powerpc/include/asm/spinlock.h index 434615f1d761..6ec72282888d 100644 --- a/arch/powerpc/include/asm/spinlock.h +++ b/arch/powerpc/include/asm/spinlock.h @@ -10,5 +10,9 @@ #include #endif +#ifndef CONFIG_PARAVIRT_SPINLOCKS +static inline void pv_spinlocks_init(void) { } +#endif + #endif /* __KERNEL__ */ #endif /* __ASM_SPINLOCK_H */ diff --git a/arch/powerpc/platforms/pseries/Kconfig b/arch/powerpc/platforms/pseries/Kconfig index 24c18362e5ea..5e037df2a3a1 100644 --- a/arch/powerpc/platforms/pseries/Kconfig +++ b/arch/powerpc/platforms/pseries/Kconfig @@ -25,15 +25,22 @@ config PPC_PSERIES select SWIOTLB default y +config PARAVIRT_SPINLOCKS + bool + config PPC_SPLPAR - depends on PPC_PSERIES bool "Support for shared-processor logical partitions" + depends on PPC_PSERIES + select PARAVIRT_SPINLOCKS if PPC_QUEUED_SPINLOCKS + default y help Enabling this option will make the kernel run more efficiently on logically-partitioned pSeries systems which use shared processors, that is, which share physical processors between two or more partitions. + Say Y if you are unsure. + config DTL bool "Dispatch Trace Log" depends on PPC_SPLPAR && DEBUG_FS diff --git a/arch/powerpc/platforms/pseries/setup.c b/arch/powerpc/platforms/pseries/setup.c index ba34eb23e8f5..e29c9bf0a3b9 100644 --- a/arch/powerpc/platforms/pseries/setup.c +++ b/arch/powerpc/platforms/pseries/setup.c @@ -773,8 +773,10 @@ static void __init pSeries_setup_arch(void) if (firmware_has_feature(FW_FEATURE_LPAR)) { vpa_init(boot_cpuid); - if (lppaca_shared_proc(get_lppaca())) + if (lppaca_shared_proc(get_lppaca())) { static_branch_enable(&shared_processor); + pv_spinlocks_init(); + } ppc_md.power_save = pseries_lpar_idle; ppc_md.enable_pmcs = pseries_lpar_enable_pmcs; diff --git a/include/asm-generic/qspinlock.h b/include/asm-generic/qspinlock.h index fb0a814d4395..38ca14e79a86 100644 --- a/include/asm-generic/qspinlock.h +++ b/include/asm-generic/qspinlock.h @@ -69,6 +69,7 @@ static __always_inline int queued_spin_trylock(struct qspinlock *lock) extern void queued_spin_lock_slowpath(struct qspinlock *lock, u32 val); +#ifndef queued_spin_lock /** * queued_spin_lock - acquire a queued spinlock * @lock: Pointer to queued spinlock structure @@ -82,6 +83,7 @@ static __always_inline void queued_spin_lock(struct qspinlock *lock) queued_spin_lock_slowpath(lock, val); } +#endif #ifndef queued_spin_unlock /** -- cgit v1.2.3