From 6c642e442e99af1ca026af55a16f23b5f8ee612a Mon Sep 17 00:00:00 2001 From: zhangdianfang Date: Fri, 30 May 2014 08:37:28 +0800 Subject: [PATCH 01/12] tools/liblockdep: Fix comparison of a boolean value with a value of 2 Comparison of a boolean value (!__init_state) with a value of 2 (done) as currently happens in the code is unlikely to succeed and causes repeated initialization of the pthread function pointers. Instead, remove boolean comparison so that we would initialize said function pointers only once. Ref: https://bugzilla.kernel.org/show_bug.cgi?id=76741 Cc: Jean Delvare Reported-by: David Binderman Signed-off-by: Dianfang Zhang Signed-off-by: Sasha Levin --- tools/lib/lockdep/preload.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/tools/lib/lockdep/preload.c b/tools/lib/lockdep/preload.c index 23bd69cb5ade..b5e52af6ddf3 100644 --- a/tools/lib/lockdep/preload.c +++ b/tools/lib/lockdep/preload.c @@ -92,7 +92,7 @@ enum { none, prepare, done, } __init_state; static void init_preload(void); static void try_init_preload(void) { - if (!__init_state != done) + if (__init_state != done) init_preload(); } From 0c37c686b336aeead2f0b982f24eafaeb19435b1 Mon Sep 17 00:00:00 2001 From: Sasha Levin Date: Mon, 7 Jul 2014 12:14:27 -0400 Subject: [PATCH 02/12] tools/liblockdep: Remove debug print left over from development Remove a debug print in init_preload() which was left over from development and isn't usefull at all currently. It was also causing false positive test results. Reported-by: S. Lockwood-Childs Signed-off-by: Sasha Levin --- tools/lib/lockdep/preload.c | 2 -- 1 file changed, 2 deletions(-) diff --git a/tools/lib/lockdep/preload.c b/tools/lib/lockdep/preload.c index b5e52af6ddf3..51ec2ec8371b 100644 --- a/tools/lib/lockdep/preload.c +++ b/tools/lib/lockdep/preload.c @@ -439,8 +439,6 @@ __attribute__((constructor)) static void init_preload(void) ll_pthread_rwlock_unlock = dlsym(RTLD_NEXT, "pthread_rwlock_unlock"); #endif - printf("%p\n", ll_pthread_mutex_trylock);fflush(stdout); - lockdep_init(); __init_state = done; From b10827814e9c81c5a14fb73c5a6e06bd85df3f94 Mon Sep 17 00:00:00 2001 From: "S. Lockwood-Childs" Date: Mon, 7 Jul 2014 00:17:33 -0700 Subject: [PATCH 03/12] tools/liblockdep: Account for bitfield changes in lockdeps lock_acquire Commit fb9edbe984 shortened held_lock->check from a 2-bit field to a 1-bit field. Make liblockdep compatible with the new definition by passing check=1 to lock_acquire() calls, rather than the old value check=2 (which inadvertently disabled checks by overflowing to 0). Without this fix, several of the test cases in liblockdep run_tests.sh were failing. Signed-off-by: S. Lockwood-Childs Signed-off-by: Sasha Levin --- tools/lib/lockdep/include/liblockdep/mutex.h | 4 ++-- tools/lib/lockdep/include/liblockdep/rwlock.h | 8 ++++---- tools/lib/lockdep/preload.c | 16 ++++++++-------- 3 files changed, 14 insertions(+), 14 deletions(-) diff --git a/tools/lib/lockdep/include/liblockdep/mutex.h b/tools/lib/lockdep/include/liblockdep/mutex.h index c342f7087147..ee53a42818ca 100644 --- a/tools/lib/lockdep/include/liblockdep/mutex.h +++ b/tools/lib/lockdep/include/liblockdep/mutex.h @@ -35,7 +35,7 @@ static inline int __mutex_init(liblockdep_pthread_mutex_t *lock, static inline int liblockdep_pthread_mutex_lock(liblockdep_pthread_mutex_t *lock) { - lock_acquire(&lock->dep_map, 0, 0, 0, 2, NULL, (unsigned long)_RET_IP_); + lock_acquire(&lock->dep_map, 0, 0, 0, 1, NULL, (unsigned long)_RET_IP_); return pthread_mutex_lock(&lock->mutex); } @@ -47,7 +47,7 @@ static inline int liblockdep_pthread_mutex_unlock(liblockdep_pthread_mutex_t *lo static inline int liblockdep_pthread_mutex_trylock(liblockdep_pthread_mutex_t *lock) { - lock_acquire(&lock->dep_map, 0, 1, 0, 2, NULL, (unsigned long)_RET_IP_); + lock_acquire(&lock->dep_map, 0, 1, 0, 1, NULL, (unsigned long)_RET_IP_); return pthread_mutex_trylock(&lock->mutex) == 0 ? 1 : 0; } diff --git a/tools/lib/lockdep/include/liblockdep/rwlock.h b/tools/lib/lockdep/include/liblockdep/rwlock.h index a680ab8c2e36..4ec03f861551 100644 --- a/tools/lib/lockdep/include/liblockdep/rwlock.h +++ b/tools/lib/lockdep/include/liblockdep/rwlock.h @@ -36,7 +36,7 @@ static inline int __rwlock_init(liblockdep_pthread_rwlock_t *lock, static inline int liblockdep_pthread_rwlock_rdlock(liblockdep_pthread_rwlock_t *lock) { - lock_acquire(&lock->dep_map, 0, 0, 2, 2, NULL, (unsigned long)_RET_IP_); + lock_acquire(&lock->dep_map, 0, 0, 2, 1, NULL, (unsigned long)_RET_IP_); return pthread_rwlock_rdlock(&lock->rwlock); } @@ -49,19 +49,19 @@ static inline int liblockdep_pthread_rwlock_unlock(liblockdep_pthread_rwlock_t * static inline int liblockdep_pthread_rwlock_wrlock(liblockdep_pthread_rwlock_t *lock) { - lock_acquire(&lock->dep_map, 0, 0, 0, 2, NULL, (unsigned long)_RET_IP_); + lock_acquire(&lock->dep_map, 0, 0, 0, 1, NULL, (unsigned long)_RET_IP_); return pthread_rwlock_wrlock(&lock->rwlock); } static inline int liblockdep_pthread_rwlock_tryrdlock(liblockdep_pthread_rwlock_t *lock) { - lock_acquire(&lock->dep_map, 0, 1, 2, 2, NULL, (unsigned long)_RET_IP_); + lock_acquire(&lock->dep_map, 0, 1, 2, 1, NULL, (unsigned long)_RET_IP_); return pthread_rwlock_tryrdlock(&lock->rwlock) == 0 ? 1 : 0; } static inline int liblockdep_pthread_rwlock_trywlock(liblockdep_pthread_rwlock_t *lock) { - lock_acquire(&lock->dep_map, 0, 1, 0, 2, NULL, (unsigned long)_RET_IP_); + lock_acquire(&lock->dep_map, 0, 1, 0, 1, NULL, (unsigned long)_RET_IP_); return pthread_rwlock_trywlock(&lock->rwlock) == 0 ? 1 : 0; } diff --git a/tools/lib/lockdep/preload.c b/tools/lib/lockdep/preload.c index 51ec2ec8371b..6f803609e498 100644 --- a/tools/lib/lockdep/preload.c +++ b/tools/lib/lockdep/preload.c @@ -252,7 +252,7 @@ int pthread_mutex_lock(pthread_mutex_t *mutex) try_init_preload(); - lock_acquire(&__get_lock(mutex)->dep_map, 0, 0, 0, 2, NULL, + lock_acquire(&__get_lock(mutex)->dep_map, 0, 0, 0, 1, NULL, (unsigned long)_RET_IP_); /* * Here's the thing with pthread mutexes: unlike the kernel variant, @@ -281,7 +281,7 @@ int pthread_mutex_trylock(pthread_mutex_t *mutex) try_init_preload(); - lock_acquire(&__get_lock(mutex)->dep_map, 0, 1, 0, 2, NULL, (unsigned long)_RET_IP_); + lock_acquire(&__get_lock(mutex)->dep_map, 0, 1, 0, 1, NULL, (unsigned long)_RET_IP_); r = ll_pthread_mutex_trylock(mutex); if (r) lock_release(&__get_lock(mutex)->dep_map, 0, (unsigned long)_RET_IP_); @@ -303,7 +303,7 @@ int pthread_mutex_unlock(pthread_mutex_t *mutex) */ r = ll_pthread_mutex_unlock(mutex); if (r) - lock_acquire(&__get_lock(mutex)->dep_map, 0, 0, 0, 2, NULL, (unsigned long)_RET_IP_); + lock_acquire(&__get_lock(mutex)->dep_map, 0, 0, 0, 1, NULL, (unsigned long)_RET_IP_); return r; } @@ -352,7 +352,7 @@ int pthread_rwlock_rdlock(pthread_rwlock_t *rwlock) init_preload(); - lock_acquire(&__get_lock(rwlock)->dep_map, 0, 0, 2, 2, NULL, (unsigned long)_RET_IP_); + lock_acquire(&__get_lock(rwlock)->dep_map, 0, 0, 2, 1, NULL, (unsigned long)_RET_IP_); r = ll_pthread_rwlock_rdlock(rwlock); if (r) lock_release(&__get_lock(rwlock)->dep_map, 0, (unsigned long)_RET_IP_); @@ -366,7 +366,7 @@ int pthread_rwlock_tryrdlock(pthread_rwlock_t *rwlock) init_preload(); - lock_acquire(&__get_lock(rwlock)->dep_map, 0, 1, 2, 2, NULL, (unsigned long)_RET_IP_); + lock_acquire(&__get_lock(rwlock)->dep_map, 0, 1, 2, 1, NULL, (unsigned long)_RET_IP_); r = ll_pthread_rwlock_tryrdlock(rwlock); if (r) lock_release(&__get_lock(rwlock)->dep_map, 0, (unsigned long)_RET_IP_); @@ -380,7 +380,7 @@ int pthread_rwlock_trywrlock(pthread_rwlock_t *rwlock) init_preload(); - lock_acquire(&__get_lock(rwlock)->dep_map, 0, 1, 0, 2, NULL, (unsigned long)_RET_IP_); + lock_acquire(&__get_lock(rwlock)->dep_map, 0, 1, 0, 1, NULL, (unsigned long)_RET_IP_); r = ll_pthread_rwlock_trywrlock(rwlock); if (r) lock_release(&__get_lock(rwlock)->dep_map, 0, (unsigned long)_RET_IP_); @@ -394,7 +394,7 @@ int pthread_rwlock_wrlock(pthread_rwlock_t *rwlock) init_preload(); - lock_acquire(&__get_lock(rwlock)->dep_map, 0, 0, 0, 2, NULL, (unsigned long)_RET_IP_); + lock_acquire(&__get_lock(rwlock)->dep_map, 0, 0, 0, 1, NULL, (unsigned long)_RET_IP_); r = ll_pthread_rwlock_wrlock(rwlock); if (r) lock_release(&__get_lock(rwlock)->dep_map, 0, (unsigned long)_RET_IP_); @@ -411,7 +411,7 @@ int pthread_rwlock_unlock(pthread_rwlock_t *rwlock) lock_release(&__get_lock(rwlock)->dep_map, 0, (unsigned long)_RET_IP_); r = ll_pthread_rwlock_unlock(rwlock); if (r) - lock_acquire(&__get_lock(rwlock)->dep_map, 0, 0, 0, 2, NULL, (unsigned long)_RET_IP_); + lock_acquire(&__get_lock(rwlock)->dep_map, 0, 0, 0, 1, NULL, (unsigned long)_RET_IP_); return r; } From 37e9562453b813d2ea527bd9531fef2c3c592847 Mon Sep 17 00:00:00 2001 From: Jason Low Date: Fri, 4 Jul 2014 20:49:32 -0700 Subject: [PATCH 04/12] locking/rwsem: Allow conservative optimistic spinning when readers have lock Commit 4fc828e24cd9 ("locking/rwsem: Support optimistic spinning") introduced a major performance regression for workloads such as xfs_repair which mix read and write locking of the mmap_sem across many threads. The result was xfs_repair ran 5x slower on 3.16-rc2 than on 3.15 and using 20x more system CPU time. Perf profiles indicate in some workloads that significant time can be spent spinning on !owner. This is because we don't set the lock owner when readers(s) obtain the rwsem. In this patch, we'll modify rwsem_can_spin_on_owner() such that we'll return false if there is no lock owner. The rationale is that if we just entered the slowpath, yet there is no lock owner, then there is a possibility that a reader has the lock. To be conservative, we'll avoid spinning in these situations. This patch reduced the total run time of the xfs_repair workload from about 4 minutes 24 seconds down to approximately 1 minute 26 seconds, back to close to the same performance as on 3.15. Retesting of AIM7, which were some of the workloads used to test the original optimistic spinning code, confirmed that we still get big performance gains with optimistic spinning, even with this additional regression fix. Davidlohr found that while the 'custom' workload took a performance hit of ~-14% to throughput for >300 users with this additional patch, the overall gain with optimistic spinning is still ~+45%. The 'disk' workload even improved by ~+15% at >1000 users. Tested-by: Dave Chinner Acked-by: Davidlohr Bueso Signed-off-by: Jason Low Signed-off-by: Peter Zijlstra Cc: Tim Chen Cc: Linus Torvalds Link: http://lkml.kernel.org/r/1404532172.2572.30.camel@j-VirtualBox Signed-off-by: Ingo Molnar --- kernel/locking/rwsem-xadd.c | 10 +++++----- 1 file changed, 5 insertions(+), 5 deletions(-) diff --git a/kernel/locking/rwsem-xadd.c b/kernel/locking/rwsem-xadd.c index dacc32142fcc..c40c7d28661d 100644 --- a/kernel/locking/rwsem-xadd.c +++ b/kernel/locking/rwsem-xadd.c @@ -285,10 +285,10 @@ static inline bool rwsem_try_write_lock_unqueued(struct rw_semaphore *sem) static inline bool rwsem_can_spin_on_owner(struct rw_semaphore *sem) { struct task_struct *owner; - bool on_cpu = true; + bool on_cpu = false; if (need_resched()) - return 0; + return false; rcu_read_lock(); owner = ACCESS_ONCE(sem->owner); @@ -297,9 +297,9 @@ static inline bool rwsem_can_spin_on_owner(struct rw_semaphore *sem) rcu_read_unlock(); /* - * If sem->owner is not set, the rwsem owner may have - * just acquired it and not set the owner yet or the rwsem - * has been released. + * If sem->owner is not set, yet we have just recently entered the + * slowpath, then there is a possibility reader(s) may have the lock. + * To be safe, avoid spinning in these situations. */ return on_cpu; } From 046a619d8e9746fa4c0e29e8c6b78e16efc008a8 Mon Sep 17 00:00:00 2001 From: Jason Low Date: Mon, 14 Jul 2014 10:27:48 -0700 Subject: [PATCH 05/12] locking/spinlocks/mcs: Rename optimistic_spin_queue() to optimistic_spin_node() Currently, the per-cpu nodes structure for the cancellable MCS spinlock is named "optimistic_spin_queue". However, in a follow up patch in the series we will be introducing a new structure that serves as the new "handle" for the lock. It would make more sense if that structure is named "optimistic_spin_queue". Additionally, since the current use of the "optimistic_spin_queue" structure are "nodes", it might be better if we rename them to "node" anyway. This preparatory patch renames all current "optimistic_spin_queue" to "optimistic_spin_node". Signed-off-by: Jason Low Signed-off-by: Peter Zijlstra Cc: Scott Norton Cc: "Paul E. McKenney" Cc: Dave Chinner Cc: Waiman Long Cc: Davidlohr Bueso Cc: Rik van Riel Cc: Andrew Morton Cc: "H. Peter Anvin" Cc: Steven Rostedt Cc: Tim Chen Cc: Konrad Rzeszutek Wilk Cc: Aswin Chandramouleeswaran Cc: Linus Torvalds Cc: Chris Mason Cc: Heiko Carstens Cc: Josef Bacik Link: http://lkml.kernel.org/r/1405358872-3732-2-git-send-email-jason.low2@hp.com Signed-off-by: Ingo Molnar --- include/linux/mutex.h | 4 ++-- include/linux/rwsem.h | 4 ++-- kernel/locking/mcs_spinlock.c | 24 ++++++++++++------------ kernel/locking/mcs_spinlock.h | 8 ++++---- 4 files changed, 20 insertions(+), 20 deletions(-) diff --git a/include/linux/mutex.h b/include/linux/mutex.h index 11692dea18aa..885f3f56a77f 100644 --- a/include/linux/mutex.h +++ b/include/linux/mutex.h @@ -46,7 +46,7 @@ * - detects multi-task circular deadlocks and prints out all affected * locks and tasks (and only those tasks) */ -struct optimistic_spin_queue; +struct optimistic_spin_node; struct mutex { /* 1: unlocked, 0: locked, negative: locked, possible waiters */ atomic_t count; @@ -56,7 +56,7 @@ struct mutex { struct task_struct *owner; #endif #ifdef CONFIG_MUTEX_SPIN_ON_OWNER - struct optimistic_spin_queue *osq; /* Spinner MCS lock */ + struct optimistic_spin_node *osq; /* Spinner MCS lock */ #endif #ifdef CONFIG_DEBUG_MUTEXES const char *name; diff --git a/include/linux/rwsem.h b/include/linux/rwsem.h index 8d79708146aa..ba3f108ddea1 100644 --- a/include/linux/rwsem.h +++ b/include/linux/rwsem.h @@ -16,7 +16,7 @@ #include -struct optimistic_spin_queue; +struct optimistic_spin_node; struct rw_semaphore; #ifdef CONFIG_RWSEM_GENERIC_SPINLOCK @@ -33,7 +33,7 @@ struct rw_semaphore { * if the owner is running on the cpu. */ struct task_struct *owner; - struct optimistic_spin_queue *osq; /* spinner MCS lock */ + struct optimistic_spin_node *osq; /* spinner MCS lock */ #endif #ifdef CONFIG_DEBUG_LOCK_ALLOC struct lockdep_map dep_map; diff --git a/kernel/locking/mcs_spinlock.c b/kernel/locking/mcs_spinlock.c index 838dc9e00669..e9866f70e828 100644 --- a/kernel/locking/mcs_spinlock.c +++ b/kernel/locking/mcs_spinlock.c @@ -14,18 +14,18 @@ * called from interrupt context and we have preemption disabled while * spinning. */ -static DEFINE_PER_CPU_SHARED_ALIGNED(struct optimistic_spin_queue, osq_node); +static DEFINE_PER_CPU_SHARED_ALIGNED(struct optimistic_spin_node, osq_node); /* * Get a stable @node->next pointer, either for unlock() or unqueue() purposes. * Can return NULL in case we were the last queued and we updated @lock instead. */ -static inline struct optimistic_spin_queue * -osq_wait_next(struct optimistic_spin_queue **lock, - struct optimistic_spin_queue *node, - struct optimistic_spin_queue *prev) +static inline struct optimistic_spin_node * +osq_wait_next(struct optimistic_spin_node **lock, + struct optimistic_spin_node *node, + struct optimistic_spin_node *prev) { - struct optimistic_spin_queue *next = NULL; + struct optimistic_spin_node *next = NULL; for (;;) { if (*lock == node && cmpxchg(lock, node, prev) == node) { @@ -59,10 +59,10 @@ osq_wait_next(struct optimistic_spin_queue **lock, return next; } -bool osq_lock(struct optimistic_spin_queue **lock) +bool osq_lock(struct optimistic_spin_node **lock) { - struct optimistic_spin_queue *node = this_cpu_ptr(&osq_node); - struct optimistic_spin_queue *prev, *next; + struct optimistic_spin_node *node = this_cpu_ptr(&osq_node); + struct optimistic_spin_node *prev, *next; node->locked = 0; node->next = NULL; @@ -149,10 +149,10 @@ unqueue: return false; } -void osq_unlock(struct optimistic_spin_queue **lock) +void osq_unlock(struct optimistic_spin_node **lock) { - struct optimistic_spin_queue *node = this_cpu_ptr(&osq_node); - struct optimistic_spin_queue *next; + struct optimistic_spin_node *node = this_cpu_ptr(&osq_node); + struct optimistic_spin_node *next; /* * Fast path for the uncontended case. diff --git a/kernel/locking/mcs_spinlock.h b/kernel/locking/mcs_spinlock.h index a2dbac4aca6b..c99dc0052f49 100644 --- a/kernel/locking/mcs_spinlock.h +++ b/kernel/locking/mcs_spinlock.h @@ -118,12 +118,12 @@ void mcs_spin_unlock(struct mcs_spinlock **lock, struct mcs_spinlock *node) * mutex_lock()/rwsem_down_{read,write}() etc. */ -struct optimistic_spin_queue { - struct optimistic_spin_queue *next, *prev; +struct optimistic_spin_node { + struct optimistic_spin_node *next, *prev; int locked; /* 1 if lock acquired */ }; -extern bool osq_lock(struct optimistic_spin_queue **lock); -extern void osq_unlock(struct optimistic_spin_queue **lock); +extern bool osq_lock(struct optimistic_spin_node **lock); +extern void osq_unlock(struct optimistic_spin_node **lock); #endif /* __LINUX_MCS_SPINLOCK_H */ From 90631822c5d307b5410500806e8ac3e63928aa3e Mon Sep 17 00:00:00 2001 From: Jason Low Date: Mon, 14 Jul 2014 10:27:49 -0700 Subject: [PATCH 06/12] locking/spinlocks/mcs: Convert osq lock to atomic_t to reduce overhead The cancellable MCS spinlock is currently used to queue threads that are doing optimistic spinning. It uses per-cpu nodes, where a thread obtaining the lock would access and queue the local node corresponding to the CPU that it's running on. Currently, the cancellable MCS lock is implemented by using pointers to these nodes. In this patch, instead of operating on pointers to the per-cpu nodes, we store the CPU numbers in which the per-cpu nodes correspond to in atomic_t. A similar concept is used with the qspinlock. By operating on the CPU # of the nodes using atomic_t instead of pointers to those nodes, this can reduce the overhead of the cancellable MCS spinlock by 32 bits (on 64 bit systems). Signed-off-by: Jason Low Signed-off-by: Peter Zijlstra Cc: Scott Norton Cc: "Paul E. McKenney" Cc: Dave Chinner Cc: Waiman Long Cc: Davidlohr Bueso Cc: Rik van Riel Cc: Andrew Morton Cc: "H. Peter Anvin" Cc: Steven Rostedt Cc: Tim Chen Cc: Konrad Rzeszutek Wilk Cc: Aswin Chandramouleeswaran Cc: Linus Torvalds Cc: Chris Mason Cc: Heiko Carstens Cc: Josef Bacik Link: http://lkml.kernel.org/r/1405358872-3732-3-git-send-email-jason.low2@hp.com Signed-off-by: Ingo Molnar --- include/linux/mutex.h | 4 +-- include/linux/osq_lock.h | 19 +++++++++++++++ include/linux/rwsem.h | 7 +++--- kernel/locking/mcs_spinlock.c | 46 +++++++++++++++++++++++++++++------ kernel/locking/mcs_spinlock.h | 5 ++-- kernel/locking/mutex.c | 2 +- kernel/locking/rwsem-xadd.c | 2 +- 7 files changed, 68 insertions(+), 17 deletions(-) create mode 100644 include/linux/osq_lock.h diff --git a/include/linux/mutex.h b/include/linux/mutex.h index 885f3f56a77f..42aa9b9ecd5f 100644 --- a/include/linux/mutex.h +++ b/include/linux/mutex.h @@ -17,6 +17,7 @@ #include #include #include +#include /* * Simple, straightforward mutexes with strict semantics: @@ -46,7 +47,6 @@ * - detects multi-task circular deadlocks and prints out all affected * locks and tasks (and only those tasks) */ -struct optimistic_spin_node; struct mutex { /* 1: unlocked, 0: locked, negative: locked, possible waiters */ atomic_t count; @@ -56,7 +56,7 @@ struct mutex { struct task_struct *owner; #endif #ifdef CONFIG_MUTEX_SPIN_ON_OWNER - struct optimistic_spin_node *osq; /* Spinner MCS lock */ + struct optimistic_spin_queue osq; /* Spinner MCS lock */ #endif #ifdef CONFIG_DEBUG_MUTEXES const char *name; diff --git a/include/linux/osq_lock.h b/include/linux/osq_lock.h new file mode 100644 index 000000000000..b001682bf7cb --- /dev/null +++ b/include/linux/osq_lock.h @@ -0,0 +1,19 @@ +#ifndef __LINUX_OSQ_LOCK_H +#define __LINUX_OSQ_LOCK_H + +/* + * An MCS like lock especially tailored for optimistic spinning for sleeping + * lock implementations (mutex, rwsem, etc). + */ + +#define OSQ_UNLOCKED_VAL (0) + +struct optimistic_spin_queue { + /* + * Stores an encoded value of the CPU # of the tail node in the queue. + * If the queue is empty, then it's set to OSQ_UNLOCKED_VAL. + */ + atomic_t tail; +}; + +#endif diff --git a/include/linux/rwsem.h b/include/linux/rwsem.h index ba3f108ddea1..9fdcdd03507d 100644 --- a/include/linux/rwsem.h +++ b/include/linux/rwsem.h @@ -13,10 +13,9 @@ #include #include #include - #include +#include -struct optimistic_spin_node; struct rw_semaphore; #ifdef CONFIG_RWSEM_GENERIC_SPINLOCK @@ -33,7 +32,7 @@ struct rw_semaphore { * if the owner is running on the cpu. */ struct task_struct *owner; - struct optimistic_spin_node *osq; /* spinner MCS lock */ + struct optimistic_spin_queue osq; /* spinner MCS lock */ #endif #ifdef CONFIG_DEBUG_LOCK_ALLOC struct lockdep_map dep_map; @@ -70,7 +69,7 @@ static inline int rwsem_is_locked(struct rw_semaphore *sem) __RAW_SPIN_LOCK_UNLOCKED(name.wait_lock), \ LIST_HEAD_INIT((name).wait_list), \ NULL, /* owner */ \ - NULL /* mcs lock */ \ + { ATOMIC_INIT(OSQ_UNLOCKED_VAL) } /* osq */ \ __RWSEM_DEP_MAP_INIT(name) } #else #define __RWSEM_INITIALIZER(name) \ diff --git a/kernel/locking/mcs_spinlock.c b/kernel/locking/mcs_spinlock.c index e9866f70e828..32fc16c0a545 100644 --- a/kernel/locking/mcs_spinlock.c +++ b/kernel/locking/mcs_spinlock.c @@ -16,19 +16,45 @@ */ static DEFINE_PER_CPU_SHARED_ALIGNED(struct optimistic_spin_node, osq_node); +/* + * We use the value 0 to represent "no CPU", thus the encoded value + * will be the CPU number incremented by 1. + */ +static inline int encode_cpu(int cpu_nr) +{ + return cpu_nr + 1; +} + +static inline struct optimistic_spin_node *decode_cpu(int encoded_cpu_val) +{ + int cpu_nr = encoded_cpu_val - 1; + + return per_cpu_ptr(&osq_node, cpu_nr); +} + /* * Get a stable @node->next pointer, either for unlock() or unqueue() purposes. * Can return NULL in case we were the last queued and we updated @lock instead. */ static inline struct optimistic_spin_node * -osq_wait_next(struct optimistic_spin_node **lock, +osq_wait_next(struct optimistic_spin_queue *lock, struct optimistic_spin_node *node, struct optimistic_spin_node *prev) { struct optimistic_spin_node *next = NULL; + int curr = encode_cpu(smp_processor_id()); + int old; + + /* + * If there is a prev node in queue, then the 'old' value will be + * the prev node's CPU #, else it's set to OSQ_UNLOCKED_VAL since if + * we're currently last in queue, then the queue will then become empty. + */ + old = prev ? prev->cpu : OSQ_UNLOCKED_VAL; for (;;) { - if (*lock == node && cmpxchg(lock, node, prev) == node) { + if (atomic_read(&lock->tail) == curr && + atomic_cmpxchg(&lock->tail, curr, old) == curr) { /* * We were the last queued, we moved @lock back. @prev * will now observe @lock and will complete its @@ -59,18 +85,23 @@ osq_wait_next(struct optimistic_spin_node **lock, return next; } -bool osq_lock(struct optimistic_spin_node **lock) +bool osq_lock(struct optimistic_spin_queue *lock) { struct optimistic_spin_node *node = this_cpu_ptr(&osq_node); struct optimistic_spin_node *prev, *next; + int curr = encode_cpu(smp_processor_id()); + int old; node->locked = 0; node->next = NULL; + node->cpu = curr; - node->prev = prev = xchg(lock, node); - if (likely(prev == NULL)) + old = atomic_xchg(&lock->tail, curr); + if (old == OSQ_UNLOCKED_VAL) return true; + prev = decode_cpu(old); + node->prev = prev; ACCESS_ONCE(prev->next) = node; /* @@ -149,15 +180,16 @@ unqueue: return false; } -void osq_unlock(struct optimistic_spin_node **lock) +void osq_unlock(struct optimistic_spin_queue *lock) { struct optimistic_spin_node *node = this_cpu_ptr(&osq_node); struct optimistic_spin_node *next; + int curr = encode_cpu(smp_processor_id()); /* * Fast path for the uncontended case. */ - if (likely(cmpxchg(lock, node, NULL) == node)) + if (likely(atomic_cmpxchg(&lock->tail, curr, OSQ_UNLOCKED_VAL) == curr)) return; /* diff --git a/kernel/locking/mcs_spinlock.h b/kernel/locking/mcs_spinlock.h index c99dc0052f49..74356dc0ce29 100644 --- a/kernel/locking/mcs_spinlock.h +++ b/kernel/locking/mcs_spinlock.h @@ -121,9 +121,10 @@ void mcs_spin_unlock(struct mcs_spinlock **lock, struct mcs_spinlock *node) struct optimistic_spin_node { struct optimistic_spin_node *next, *prev; int locked; /* 1 if lock acquired */ + int cpu; /* encoded CPU # value */ }; -extern bool osq_lock(struct optimistic_spin_node **lock); -extern void osq_unlock(struct optimistic_spin_node **lock); +extern bool osq_lock(struct optimistic_spin_queue *lock); +extern void osq_unlock(struct optimistic_spin_queue *lock); #endif /* __LINUX_MCS_SPINLOCK_H */ diff --git a/kernel/locking/mutex.c b/kernel/locking/mutex.c index bc73d33c6760..d9b313906caa 100644 --- a/kernel/locking/mutex.c +++ b/kernel/locking/mutex.c @@ -60,7 +60,7 @@ __mutex_init(struct mutex *lock, const char *name, struct lock_class_key *key) INIT_LIST_HEAD(&lock->wait_list); mutex_clear_owner(lock); #ifdef CONFIG_MUTEX_SPIN_ON_OWNER - lock->osq = NULL; + atomic_set(&lock->osq.tail, OSQ_UNLOCKED_VAL); #endif debug_mutex_init(lock, name, key); diff --git a/kernel/locking/rwsem-xadd.c b/kernel/locking/rwsem-xadd.c index c40c7d28661d..b77a6230bbf6 100644 --- a/kernel/locking/rwsem-xadd.c +++ b/kernel/locking/rwsem-xadd.c @@ -84,7 +84,7 @@ void __init_rwsem(struct rw_semaphore *sem, const char *name, INIT_LIST_HEAD(&sem->wait_list); #ifdef CONFIG_SMP sem->owner = NULL; - sem->osq = NULL; + atomic_set(&sem->osq.tail, OSQ_UNLOCKED_VAL); #endif } From 4d9d951e6b5df85ccfca2c5bd8b4f5c71d256b65 Mon Sep 17 00:00:00 2001 From: Jason Low Date: Mon, 14 Jul 2014 10:27:50 -0700 Subject: [PATCH 07/12] locking/spinlocks/mcs: Introduce and use init macro and function for osq locks Currently, we initialize the osq lock by directly setting the lock's values. It would be preferable if we use an init macro to do the initialization like we do with other locks. This patch introduces and uses a macro and function for initializing the osq lock. Signed-off-by: Jason Low Signed-off-by: Peter Zijlstra Cc: Scott Norton Cc: "Paul E. McKenney" Cc: Dave Chinner Cc: Waiman Long Cc: Davidlohr Bueso Cc: Rik van Riel Cc: Andrew Morton Cc: "H. Peter Anvin" Cc: Steven Rostedt Cc: Tim Chen Cc: Konrad Rzeszutek Wilk Cc: Aswin Chandramouleeswaran Cc: Linus Torvalds Cc: Chris Mason Cc: Josef Bacik Link: http://lkml.kernel.org/r/1405358872-3732-4-git-send-email-jason.low2@hp.com Signed-off-by: Ingo Molnar --- include/linux/osq_lock.h | 8 ++++++++ include/linux/rwsem.h | 2 +- kernel/locking/mutex.c | 2 +- kernel/locking/rwsem-xadd.c | 2 +- 4 files changed, 11 insertions(+), 3 deletions(-) diff --git a/include/linux/osq_lock.h b/include/linux/osq_lock.h index b001682bf7cb..90230d5811c5 100644 --- a/include/linux/osq_lock.h +++ b/include/linux/osq_lock.h @@ -16,4 +16,12 @@ struct optimistic_spin_queue { atomic_t tail; }; +/* Init macro and function. */ +#define OSQ_LOCK_UNLOCKED { ATOMIC_INIT(OSQ_UNLOCKED_VAL) } + +static inline void osq_lock_init(struct optimistic_spin_queue *lock) +{ + atomic_set(&lock->tail, OSQ_UNLOCKED_VAL); +} + #endif diff --git a/include/linux/rwsem.h b/include/linux/rwsem.h index 9fdcdd03507d..25cd9aa2f3d7 100644 --- a/include/linux/rwsem.h +++ b/include/linux/rwsem.h @@ -69,7 +69,7 @@ static inline int rwsem_is_locked(struct rw_semaphore *sem) __RAW_SPIN_LOCK_UNLOCKED(name.wait_lock), \ LIST_HEAD_INIT((name).wait_list), \ NULL, /* owner */ \ - { ATOMIC_INIT(OSQ_UNLOCKED_VAL) } /* osq */ \ + OSQ_LOCK_UNLOCKED /* osq */ \ __RWSEM_DEP_MAP_INIT(name) } #else #define __RWSEM_INITIALIZER(name) \ diff --git a/kernel/locking/mutex.c b/kernel/locking/mutex.c index d9b313906caa..acca2c1a3c5e 100644 --- a/kernel/locking/mutex.c +++ b/kernel/locking/mutex.c @@ -60,7 +60,7 @@ __mutex_init(struct mutex *lock, const char *name, struct lock_class_key *key) INIT_LIST_HEAD(&lock->wait_list); mutex_clear_owner(lock); #ifdef CONFIG_MUTEX_SPIN_ON_OWNER - atomic_set(&lock->osq.tail, OSQ_UNLOCKED_VAL); + osq_lock_init(&lock->osq); #endif debug_mutex_init(lock, name, key); diff --git a/kernel/locking/rwsem-xadd.c b/kernel/locking/rwsem-xadd.c index b77a6230bbf6..7190592c2645 100644 --- a/kernel/locking/rwsem-xadd.c +++ b/kernel/locking/rwsem-xadd.c @@ -84,7 +84,7 @@ void __init_rwsem(struct rw_semaphore *sem, const char *name, INIT_LIST_HEAD(&sem->wait_list); #ifdef CONFIG_SMP sem->owner = NULL; - atomic_set(&sem->osq.tail, OSQ_UNLOCKED_VAL); + osq_lock_init(&sem->osq); #endif } From 33ecd2083a9560fbc1ef1b1279ef3ecb4c012a4f Mon Sep 17 00:00:00 2001 From: Jason Low Date: Mon, 14 Jul 2014 10:27:51 -0700 Subject: [PATCH 08/12] locking/spinlocks/mcs: Micro-optimize osq_unlock() In the unlock function of the cancellable MCS spinlock, the first thing we do is to retrive the current CPU's osq node. However, due to the changes made in the previous patch, in the common case where the lock is not contended, we wouldn't need to access the current CPU's osq node anymore. This patch optimizes this by only retriving this CPU's osq node after we attempt the initial cmpxchg to unlock the osq and found that its contended. Signed-off-by: Jason Low Signed-off-by: Peter Zijlstra Cc: Scott Norton Cc: "Paul E. McKenney" Cc: Dave Chinner Cc: Waiman Long Cc: Davidlohr Bueso Cc: Rik van Riel Cc: Andrew Morton Cc: "H. Peter Anvin" Cc: Steven Rostedt Cc: Tim Chen Cc: Konrad Rzeszutek Wilk Cc: Aswin Chandramouleeswaran Cc: Linus Torvalds Link: http://lkml.kernel.org/r/1405358872-3732-5-git-send-email-jason.low2@hp.com Signed-off-by: Ingo Molnar --- kernel/locking/mcs_spinlock.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/kernel/locking/mcs_spinlock.c b/kernel/locking/mcs_spinlock.c index 32fc16c0a545..be9ee1559fca 100644 --- a/kernel/locking/mcs_spinlock.c +++ b/kernel/locking/mcs_spinlock.c @@ -182,8 +182,7 @@ unqueue: void osq_unlock(struct optimistic_spin_queue *lock) { - struct optimistic_spin_node *node = this_cpu_ptr(&osq_node); - struct optimistic_spin_node *next; + struct optimistic_spin_node *node, *next; int curr = encode_cpu(smp_processor_id()); /* @@ -195,6 +194,7 @@ void osq_unlock(struct optimistic_spin_queue *lock) /* * Second most likely case. */ + node = this_cpu_ptr(&osq_node); next = xchg(&node->next, NULL); if (next) { ACCESS_ONCE(next->locked) = 1; From 13b9a962a2594ee880c5d50d7f70964da1d4fe5a Mon Sep 17 00:00:00 2001 From: Peter Zijlstra Date: Wed, 16 Jul 2014 14:54:55 +0200 Subject: [PATCH 09/12] locking/rwsem: Rename 'activity' to 'count' There are two definitions of struct rw_semaphore, one in linux/rwsem.h and one in linux/rwsem-spinlock.h. For some reason they have different names for the initial field. This makes it impossible to use C99 named initialization for __RWSEM_INITIALIZER() -- or we have to duplicate that entire thing along with the structure definitions. The simpler patch is renaming the rwsem-spinlock variant to match the regular rwsem. This allows us to switch to C99 named initialization. Signed-off-by: Peter Zijlstra Cc: Linus Torvalds Link: http://lkml.kernel.org/n/tip-bmrZolsbGmautmzrerog27io@git.kernel.org Signed-off-by: Ingo Molnar --- include/linux/rwsem-spinlock.h | 8 ++++---- kernel/locking/rwsem-spinlock.c | 28 ++++++++++++++-------------- 2 files changed, 18 insertions(+), 18 deletions(-) diff --git a/include/linux/rwsem-spinlock.h b/include/linux/rwsem-spinlock.h index d5b13bc07a0b..561e8615528d 100644 --- a/include/linux/rwsem-spinlock.h +++ b/include/linux/rwsem-spinlock.h @@ -15,13 +15,13 @@ #ifdef __KERNEL__ /* * the rw-semaphore definition - * - if activity is 0 then there are no active readers or writers - * - if activity is +ve then that is the number of active readers - * - if activity is -1 then there is one active writer + * - if count is 0 then there are no active readers or writers + * - if count is +ve then that is the number of active readers + * - if count is -1 then there is one active writer * - if wait_list is not empty, then there are processes waiting for the semaphore */ struct rw_semaphore { - __s32 activity; + __s32 count; raw_spinlock_t wait_lock; struct list_head wait_list; #ifdef CONFIG_DEBUG_LOCK_ALLOC diff --git a/kernel/locking/rwsem-spinlock.c b/kernel/locking/rwsem-spinlock.c index 9be8a9144978..2c93571162cb 100644 --- a/kernel/locking/rwsem-spinlock.c +++ b/kernel/locking/rwsem-spinlock.c @@ -26,7 +26,7 @@ int rwsem_is_locked(struct rw_semaphore *sem) unsigned long flags; if (raw_spin_trylock_irqsave(&sem->wait_lock, flags)) { - ret = (sem->activity != 0); + ret = (sem->count != 0); raw_spin_unlock_irqrestore(&sem->wait_lock, flags); } return ret; @@ -46,7 +46,7 @@ void __init_rwsem(struct rw_semaphore *sem, const char *name, debug_check_no_locks_freed((void *)sem, sizeof(*sem)); lockdep_init_map(&sem->dep_map, name, key, 0); #endif - sem->activity = 0; + sem->count = 0; raw_spin_lock_init(&sem->wait_lock); INIT_LIST_HEAD(&sem->wait_list); } @@ -95,7 +95,7 @@ __rwsem_do_wake(struct rw_semaphore *sem, int wakewrite) waiter = list_entry(next, struct rwsem_waiter, list); } while (waiter->type != RWSEM_WAITING_FOR_WRITE); - sem->activity += woken; + sem->count += woken; out: return sem; @@ -126,9 +126,9 @@ void __sched __down_read(struct rw_semaphore *sem) raw_spin_lock_irqsave(&sem->wait_lock, flags); - if (sem->activity >= 0 && list_empty(&sem->wait_list)) { + if (sem->count >= 0 && list_empty(&sem->wait_list)) { /* granted */ - sem->activity++; + sem->count++; raw_spin_unlock_irqrestore(&sem->wait_lock, flags); goto out; } @@ -170,9 +170,9 @@ int __down_read_trylock(struct rw_semaphore *sem) raw_spin_lock_irqsave(&sem->wait_lock, flags); - if (sem->activity >= 0 && list_empty(&sem->wait_list)) { + if (sem->count >= 0 && list_empty(&sem->wait_list)) { /* granted */ - sem->activity++; + sem->count++; ret = 1; } @@ -206,7 +206,7 @@ void __sched __down_write_nested(struct rw_semaphore *sem, int subclass) * itself into sleep and waiting for system woke it or someone * else in the head of the wait list up. */ - if (sem->activity == 0) + if (sem->count == 0) break; set_task_state(tsk, TASK_UNINTERRUPTIBLE); raw_spin_unlock_irqrestore(&sem->wait_lock, flags); @@ -214,7 +214,7 @@ void __sched __down_write_nested(struct rw_semaphore *sem, int subclass) raw_spin_lock_irqsave(&sem->wait_lock, flags); } /* got the lock */ - sem->activity = -1; + sem->count = -1; list_del(&waiter.list); raw_spin_unlock_irqrestore(&sem->wait_lock, flags); @@ -235,9 +235,9 @@ int __down_write_trylock(struct rw_semaphore *sem) raw_spin_lock_irqsave(&sem->wait_lock, flags); - if (sem->activity == 0) { + if (sem->count == 0) { /* got the lock */ - sem->activity = -1; + sem->count = -1; ret = 1; } @@ -255,7 +255,7 @@ void __up_read(struct rw_semaphore *sem) raw_spin_lock_irqsave(&sem->wait_lock, flags); - if (--sem->activity == 0 && !list_empty(&sem->wait_list)) + if (--sem->count == 0 && !list_empty(&sem->wait_list)) sem = __rwsem_wake_one_writer(sem); raw_spin_unlock_irqrestore(&sem->wait_lock, flags); @@ -270,7 +270,7 @@ void __up_write(struct rw_semaphore *sem) raw_spin_lock_irqsave(&sem->wait_lock, flags); - sem->activity = 0; + sem->count = 0; if (!list_empty(&sem->wait_list)) sem = __rwsem_do_wake(sem, 1); @@ -287,7 +287,7 @@ void __downgrade_write(struct rw_semaphore *sem) raw_spin_lock_irqsave(&sem->wait_lock, flags); - sem->activity = 1; + sem->count = 1; if (!list_empty(&sem->wait_list)) sem = __rwsem_do_wake(sem, 0); From ce069fc920e5734558b3d9cbef1ab06cf01ee793 Mon Sep 17 00:00:00 2001 From: Jason Low Date: Mon, 14 Jul 2014 10:27:52 -0700 Subject: [PATCH 10/12] locking/rwsem: Reduce the size of struct rw_semaphore Recent optimistic spinning additions to rwsem provide significant performance benefits on many workloads on large machines. The cost of it was increasing the size of the rwsem structure by up to 128 bits. However, now that the previous patches in this series bring the overhead of struct optimistic_spin_queue to 32 bits, this patch reorders some fields in struct rw_semaphore such that we can reduce the overhead of the rwsem structure by 64 bits (on 64 bit systems). The extra overhead required for rwsem optimistic spinning would now be up to 8 additional bytes instead of up to 16 bytes. Additionally, the size of rwsem would now be more in line with mutexes. Signed-off-by: Jason Low Signed-off-by: Peter Zijlstra Cc: Scott Norton Cc: "Paul E. McKenney" Cc: Dave Chinner Cc: Waiman Long Cc: Davidlohr Bueso Cc: Rik van Riel Cc: Andrew Morton Cc: "H. Peter Anvin" Cc: Steven Rostedt Cc: Tim Chen Cc: Konrad Rzeszutek Wilk Cc: Aswin Chandramouleeswaran Cc: Linus Torvalds Cc: Chris Mason Cc: Josef Bacik Link: http://lkml.kernel.org/r/1405358872-3732-6-git-send-email-jason.low2@hp.com Signed-off-by: Ingo Molnar --- include/linux/rwsem.h | 25 +++++++++++-------------- 1 file changed, 11 insertions(+), 14 deletions(-) diff --git a/include/linux/rwsem.h b/include/linux/rwsem.h index 25cd9aa2f3d7..716807f0eb2d 100644 --- a/include/linux/rwsem.h +++ b/include/linux/rwsem.h @@ -24,15 +24,15 @@ struct rw_semaphore; /* All arch specific implementations share the same struct */ struct rw_semaphore { long count; - raw_spinlock_t wait_lock; struct list_head wait_list; + raw_spinlock_t wait_lock; #ifdef CONFIG_SMP + struct optimistic_spin_queue osq; /* spinner MCS lock */ /* * Write owner. Used as a speculative check to see * if the owner is running on the cpu. */ struct task_struct *owner; - struct optimistic_spin_queue osq; /* spinner MCS lock */ #endif #ifdef CONFIG_DEBUG_LOCK_ALLOC struct lockdep_map dep_map; @@ -64,21 +64,18 @@ static inline int rwsem_is_locked(struct rw_semaphore *sem) #endif #if defined(CONFIG_SMP) && !defined(CONFIG_RWSEM_GENERIC_SPINLOCK) -#define __RWSEM_INITIALIZER(name) \ - { RWSEM_UNLOCKED_VALUE, \ - __RAW_SPIN_LOCK_UNLOCKED(name.wait_lock), \ - LIST_HEAD_INIT((name).wait_list), \ - NULL, /* owner */ \ - OSQ_LOCK_UNLOCKED /* osq */ \ - __RWSEM_DEP_MAP_INIT(name) } +#define __RWSEM_OPT_INIT(lockname) , .osq = OSQ_LOCK_UNLOCKED, .owner = NULL #else -#define __RWSEM_INITIALIZER(name) \ - { RWSEM_UNLOCKED_VALUE, \ - __RAW_SPIN_LOCK_UNLOCKED(name.wait_lock), \ - LIST_HEAD_INIT((name).wait_list) \ - __RWSEM_DEP_MAP_INIT(name) } +#define __RWSEM_OPT_INIT(lockname) #endif +#define __RWSEM_INITIALIZER(name) \ + { .count = RWSEM_UNLOCKED_VALUE, \ + .wait_list = LIST_HEAD_INIT((name).wait_list), \ + .wait_lock = __RAW_SPIN_LOCK_UNLOCKED(name.wait_lock) \ + __RWSEM_OPT_INIT(name) \ + __RWSEM_DEP_MAP_INIT(name) } + #define DECLARE_RWSEM(name) \ struct rw_semaphore name = __RWSEM_INITIALIZER(name) From 4badad352a6bb202ec68afa7a574c0bb961e5ebc Mon Sep 17 00:00:00 2001 From: Peter Zijlstra Date: Fri, 6 Jun 2014 19:53:16 +0200 Subject: [PATCH 11/12] locking/mutex: Disable optimistic spinning on some architectures The optimistic spin code assumes regular stores and cmpxchg() play nice; this is found to not be true for at least: parisc, sparc32, tile32, metag-lock1, arc-!llsc and hexagon. There is further wreckage, but this in particular seemed easy to trigger, so blacklist this. Opt in for known good archs. Signed-off-by: Peter Zijlstra Reported-by: Mikulas Patocka Cc: David Miller Cc: Chris Metcalf Cc: James Bottomley Cc: Vineet Gupta Cc: Jason Low Cc: Waiman Long Cc: "James E.J. Bottomley" Cc: Paul McKenney Cc: John David Anglin Cc: James Hogan Cc: Linus Torvalds Cc: Davidlohr Bueso Cc: stable@vger.kernel.org Cc: Benjamin Herrenschmidt Cc: Catalin Marinas Cc: Russell King Cc: Will Deacon Cc: linux-arm-kernel@lists.infradead.org Cc: linux-kernel@vger.kernel.org Cc: linuxppc-dev@lists.ozlabs.org Cc: sparclinux@vger.kernel.org Link: http://lkml.kernel.org/r/20140606175316.GV13930@laptop.programming.kicks-ass.net Signed-off-by: Ingo Molnar --- arch/arm/Kconfig | 1 + arch/arm64/Kconfig | 1 + arch/powerpc/Kconfig | 1 + arch/sparc/Kconfig | 1 + arch/x86/Kconfig | 1 + kernel/Kconfig.locks | 5 ++++- 6 files changed, 9 insertions(+), 1 deletion(-) diff --git a/arch/arm/Kconfig b/arch/arm/Kconfig index 245058b3b0ef..88acf8bc1490 100644 --- a/arch/arm/Kconfig +++ b/arch/arm/Kconfig @@ -6,6 +6,7 @@ config ARM select ARCH_HAS_TICK_BROADCAST if GENERIC_CLOCKEVENTS_BROADCAST select ARCH_HAVE_CUSTOM_GPIO_H select ARCH_MIGHT_HAVE_PC_PARPORT + select ARCH_SUPPORTS_ATOMIC_RMW select ARCH_USE_BUILTIN_BSWAP select ARCH_USE_CMPXCHG_LOCKREF select ARCH_WANT_IPC_PARSE_VERSION diff --git a/arch/arm64/Kconfig b/arch/arm64/Kconfig index a474de346be6..839f48c26ef0 100644 --- a/arch/arm64/Kconfig +++ b/arch/arm64/Kconfig @@ -4,6 +4,7 @@ config ARM64 select ARCH_HAS_OPP select ARCH_HAS_TICK_BROADCAST if GENERIC_CLOCKEVENTS_BROADCAST select ARCH_USE_CMPXCHG_LOCKREF + select ARCH_SUPPORTS_ATOMIC_RMW select ARCH_WANT_OPTIONAL_GPIOLIB select ARCH_WANT_COMPAT_IPC_PARSE_VERSION select ARCH_WANT_FRAME_POINTERS diff --git a/arch/powerpc/Kconfig b/arch/powerpc/Kconfig index fefe7c8bf05f..80b94b0add1f 100644 --- a/arch/powerpc/Kconfig +++ b/arch/powerpc/Kconfig @@ -145,6 +145,7 @@ config PPC select HAVE_IRQ_EXIT_ON_IRQ_STACK select ARCH_USE_CMPXCHG_LOCKREF if PPC64 select HAVE_ARCH_AUDITSYSCALL + select ARCH_SUPPORTS_ATOMIC_RMW config GENERIC_CSUM def_bool CPU_LITTLE_ENDIAN diff --git a/arch/sparc/Kconfig b/arch/sparc/Kconfig index 29f2e988c56a..407c87d9879a 100644 --- a/arch/sparc/Kconfig +++ b/arch/sparc/Kconfig @@ -78,6 +78,7 @@ config SPARC64 select HAVE_C_RECORDMCOUNT select NO_BOOTMEM select HAVE_ARCH_AUDITSYSCALL + select ARCH_SUPPORTS_ATOMIC_RMW config ARCH_DEFCONFIG string diff --git a/arch/x86/Kconfig b/arch/x86/Kconfig index a8f749ef0fdc..d24887b645dc 100644 --- a/arch/x86/Kconfig +++ b/arch/x86/Kconfig @@ -131,6 +131,7 @@ config X86 select HAVE_CC_STACKPROTECTOR select GENERIC_CPU_AUTOPROBE select HAVE_ARCH_AUDITSYSCALL + select ARCH_SUPPORTS_ATOMIC_RMW config INSTRUCTION_DECODER def_bool y diff --git a/kernel/Kconfig.locks b/kernel/Kconfig.locks index 35536d9c0964..81907941d921 100644 --- a/kernel/Kconfig.locks +++ b/kernel/Kconfig.locks @@ -220,9 +220,12 @@ config INLINE_WRITE_UNLOCK_IRQRESTORE endif +config ARCH_SUPPORTS_ATOMIC_RMW + bool + config MUTEX_SPIN_ON_OWNER def_bool y - depends on SMP && !DEBUG_MUTEXES + depends on SMP && !DEBUG_MUTEXES && ARCH_SUPPORTS_ATOMIC_RMW config ARCH_USE_QUEUE_RWLOCK bool From 5db6c6fefb1ca0e81e3bd6dd8998bf51c453d823 Mon Sep 17 00:00:00 2001 From: Davidlohr Bueso Date: Fri, 11 Jul 2014 14:00:06 -0700 Subject: [PATCH 12/12] locking/rwsem: Add CONFIG_RWSEM_SPIN_ON_OWNER Just like with mutexes (CONFIG_MUTEX_SPIN_ON_OWNER), encapsulate the dependencies for rwsem optimistic spinning. No logical changes here as it continues to depend on both SMP and the XADD algorithm variant. Signed-off-by: Davidlohr Bueso Acked-by: Jason Low [ Also make it depend on ARCH_SUPPORTS_ATOMIC_RMW. ] Signed-off-by: Peter Zijlstra Link: http://lkml.kernel.org/r/1405112406-13052-2-git-send-email-davidlohr@hp.com Cc: aswin@hp.com Cc: Chris Mason Cc: Davidlohr Bueso Cc: Josef Bacik Cc: Linus Torvalds Cc: Waiman Long Signed-off-by: Ingo Molnar Signed-off-by: Ingo Molnar --- include/linux/rwsem.h | 6 ++++-- kernel/Kconfig.locks | 4 ++++ kernel/locking/rwsem-xadd.c | 4 ++-- kernel/locking/rwsem.c | 2 +- 4 files changed, 11 insertions(+), 5 deletions(-) diff --git a/include/linux/rwsem.h b/include/linux/rwsem.h index 716807f0eb2d..035d3c57fc8a 100644 --- a/include/linux/rwsem.h +++ b/include/linux/rwsem.h @@ -14,7 +14,9 @@ #include #include #include +#ifdef CONFIG_RWSEM_SPIN_ON_OWNER #include +#endif struct rw_semaphore; @@ -26,7 +28,7 @@ struct rw_semaphore { long count; struct list_head wait_list; raw_spinlock_t wait_lock; -#ifdef CONFIG_SMP +#ifdef CONFIG_RWSEM_SPIN_ON_OWNER struct optimistic_spin_queue osq; /* spinner MCS lock */ /* * Write owner. Used as a speculative check to see @@ -63,7 +65,7 @@ static inline int rwsem_is_locked(struct rw_semaphore *sem) # define __RWSEM_DEP_MAP_INIT(lockname) #endif -#if defined(CONFIG_SMP) && !defined(CONFIG_RWSEM_GENERIC_SPINLOCK) +#ifdef CONFIG_RWSEM_SPIN_ON_OWNER #define __RWSEM_OPT_INIT(lockname) , .osq = OSQ_LOCK_UNLOCKED, .owner = NULL #else #define __RWSEM_OPT_INIT(lockname) diff --git a/kernel/Kconfig.locks b/kernel/Kconfig.locks index 81907941d921..76768ee812b2 100644 --- a/kernel/Kconfig.locks +++ b/kernel/Kconfig.locks @@ -227,6 +227,10 @@ config MUTEX_SPIN_ON_OWNER def_bool y depends on SMP && !DEBUG_MUTEXES && ARCH_SUPPORTS_ATOMIC_RMW +config RWSEM_SPIN_ON_OWNER + def_bool y + depends on SMP && RWSEM_XCHGADD_ALGORITHM && ARCH_SUPPORTS_ATOMIC_RMW + config ARCH_USE_QUEUE_RWLOCK bool diff --git a/kernel/locking/rwsem-xadd.c b/kernel/locking/rwsem-xadd.c index 7190592c2645..a2391ac135c8 100644 --- a/kernel/locking/rwsem-xadd.c +++ b/kernel/locking/rwsem-xadd.c @@ -82,7 +82,7 @@ void __init_rwsem(struct rw_semaphore *sem, const char *name, sem->count = RWSEM_UNLOCKED_VALUE; raw_spin_lock_init(&sem->wait_lock); INIT_LIST_HEAD(&sem->wait_list); -#ifdef CONFIG_SMP +#ifdef CONFIG_RWSEM_SPIN_ON_OWNER sem->owner = NULL; osq_lock_init(&sem->osq); #endif @@ -262,7 +262,7 @@ static inline bool rwsem_try_write_lock(long count, struct rw_semaphore *sem) return false; } -#ifdef CONFIG_SMP +#ifdef CONFIG_RWSEM_SPIN_ON_OWNER /* * Try to acquire write lock before the writer has been put on wait queue. */ diff --git a/kernel/locking/rwsem.c b/kernel/locking/rwsem.c index 42f806de49d4..e2d3bc7f03b4 100644 --- a/kernel/locking/rwsem.c +++ b/kernel/locking/rwsem.c @@ -12,7 +12,7 @@ #include -#if defined(CONFIG_SMP) && defined(CONFIG_RWSEM_XCHGADD_ALGORITHM) +#ifdef CONFIG_RWSEM_SPIN_ON_OWNER static inline void rwsem_set_owner(struct rw_semaphore *sem) { sem->owner = current;