4 * Mutexes: blocking mutual exclusion locks
6 * Started by Ingo Molnar:
8 * Copyright (C) 2004, 2005, 2006 Red Hat, Inc., Ingo Molnar <mingo@redhat.com>
10 * Many thanks to Arjan van de Ven, Thomas Gleixner, Steven Rostedt and
11 * David Howells for suggestions and improvements.
13 * - Adaptive spinning for mutexes by Peter Zijlstra. (Ported to mainline
14 * from the -rt tree, where it was originally implemented for rtmutexes
15 * by Steven Rostedt, based on work by Gregory Haskins, Peter Morreale
18 * Also see Documentation/mutex-design.txt.
20 #include <linux/mutex.h>
21 #include <linux/sched.h>
22 #include <linux/sched/rt.h>
23 #include <linux/export.h>
24 #include <linux/spinlock.h>
25 #include <linux/interrupt.h>
26 #include <linux/debug_locks.h>
29 * In the DEBUG case we are using the "NULL fastpath" for mutexes,
30 * which forces all calls into the slowpath:
32 #ifdef CONFIG_DEBUG_MUTEXES
33 # include "mutex-debug.h"
34 # include <asm-generic/mutex-null.h>
37 # include <asm/mutex.h>
41 * A negative mutex count indicates that waiters are sleeping waiting for the
44 #define MUTEX_SHOW_NO_WAITER(mutex) (atomic_read(&(mutex)->count) >= 0)
47 __mutex_init(struct mutex *lock, const char *name, struct lock_class_key *key)
49 atomic_set(&lock->count, 1);
50 spin_lock_init(&lock->wait_lock);
51 INIT_LIST_HEAD(&lock->wait_list);
52 mutex_clear_owner(lock);
53 #ifdef CONFIG_MUTEX_SPIN_ON_OWNER
54 lock->spin_mlock = NULL;
57 debug_mutex_init(lock, name, key);
60 EXPORT_SYMBOL(__mutex_init);
62 #ifndef CONFIG_DEBUG_LOCK_ALLOC
64 * We split the mutex lock/unlock logic into separate fastpath and
65 * slowpath functions, to reduce the register pressure on the fastpath.
66 * We also put the fastpath first in the kernel image, to make sure the
67 * branch is predicted by the CPU as default-untaken.
69 static __used noinline void __sched
70 __mutex_lock_slowpath(atomic_t *lock_count);
73 * mutex_lock - acquire the mutex
74 * @lock: the mutex to be acquired
76 * Lock the mutex exclusively for this task. If the mutex is not
77 * available right now, it will sleep until it can get it.
79 * The mutex must later on be released by the same task that
80 * acquired it. Recursive locking is not allowed. The task
81 * may not exit without first unlocking the mutex. Also, kernel
82 * memory where the mutex resides mutex must not be freed with
83 * the mutex still locked. The mutex must first be initialized
84 * (or statically defined) before it can be locked. memset()-ing
85 * the mutex to 0 is not allowed.
87 * ( The CONFIG_DEBUG_MUTEXES .config option turns on debugging
88 * checks that will enforce the restrictions and will also do
89 * deadlock debugging. )
91 * This function is similar to (but not equivalent to) down().
93 void __sched mutex_lock(struct mutex *lock)
97 * The locking fastpath is the 1->0 transition from
98 * 'unlocked' into 'locked' state.
100 __mutex_fastpath_lock(&lock->count, __mutex_lock_slowpath);
101 mutex_set_owner(lock);
104 EXPORT_SYMBOL(mutex_lock);
107 #ifdef CONFIG_MUTEX_SPIN_ON_OWNER
109 * In order to avoid a stampede of mutex spinners from acquiring the mutex
110 * more or less simultaneously, the spinners need to acquire a MCS lock
111 * first before spinning on the owner field.
113 * We don't inline mspin_lock() so that perf can correctly account for the
114 * time spent in this lock function.
117 struct mspin_node *next ;
118 int locked; /* 1 if lock acquired */
120 #define MLOCK(mutex) ((struct mspin_node **)&((mutex)->spin_mlock))
123 void mspin_lock(struct mspin_node **lock, struct mspin_node *node)
125 struct mspin_node *prev;
131 prev = xchg(lock, node);
132 if (likely(prev == NULL)) {
137 ACCESS_ONCE(prev->next) = node;
139 /* Wait until the lock holder passes the lock down */
140 while (!ACCESS_ONCE(node->locked))
141 arch_mutex_cpu_relax();
144 static void mspin_unlock(struct mspin_node **lock, struct mspin_node *node)
146 struct mspin_node *next = ACCESS_ONCE(node->next);
150 * Release the lock by setting it to NULL
152 if (cmpxchg(lock, node, NULL) == node)
154 /* Wait until the next pointer is set */
155 while (!(next = ACCESS_ONCE(node->next)))
156 arch_mutex_cpu_relax();
158 ACCESS_ONCE(next->locked) = 1;
163 * Mutex spinning code migrated from kernel/sched/core.c
166 static inline bool owner_running(struct mutex *lock, struct task_struct *owner)
168 if (lock->owner != owner)
172 * Ensure we emit the owner->on_cpu, dereference _after_ checking
173 * lock->owner still matches owner, if that fails, owner might
174 * point to free()d memory, if it still matches, the rcu_read_lock()
175 * ensures the memory stays valid.
179 return owner->on_cpu;
183 * Look out! "owner" is an entirely speculative pointer
184 * access and not reliable.
187 int mutex_spin_on_owner(struct mutex *lock, struct task_struct *owner)
190 while (owner_running(lock, owner)) {
194 arch_mutex_cpu_relax();
199 * We break out the loop above on need_resched() and when the
200 * owner changed, which is a sign for heavy contention. Return
201 * success only when lock->owner is NULL.
203 return lock->owner == NULL;
207 * Initial check for entering the mutex spinning loop
209 static inline int mutex_can_spin_on_owner(struct mutex *lock)
215 retval = lock->owner->on_cpu;
218 * if lock->owner is not set, the mutex owner may have just acquired
219 * it and not set the owner yet or the mutex has been released.
225 static __used noinline void __sched __mutex_unlock_slowpath(atomic_t *lock_count);
228 * mutex_unlock - release the mutex
229 * @lock: the mutex to be released
231 * Unlock a mutex that has been locked by this task previously.
233 * This function must not be used in interrupt context. Unlocking
234 * of a not locked mutex is not allowed.
236 * This function is similar to (but not equivalent to) up().
238 void __sched mutex_unlock(struct mutex *lock)
241 * The unlocking fastpath is the 0->1 transition from 'locked'
242 * into 'unlocked' state:
244 #ifndef CONFIG_DEBUG_MUTEXES
246 * When debugging is enabled we must not clear the owner before time,
247 * the slow path will always be taken, and that clears the owner field
248 * after verifying that it was indeed current.
250 mutex_clear_owner(lock);
252 __mutex_fastpath_unlock(&lock->count, __mutex_unlock_slowpath);
255 EXPORT_SYMBOL(mutex_unlock);
258 * Lock a mutex (possibly interruptible), slowpath:
260 static inline int __sched
261 __mutex_lock_common(struct mutex *lock, long state, unsigned int subclass,
262 struct lockdep_map *nest_lock, unsigned long ip)
264 struct task_struct *task = current;
265 struct mutex_waiter waiter;
269 mutex_acquire_nest(&lock->dep_map, subclass, 0, nest_lock, ip);
271 #ifdef CONFIG_MUTEX_SPIN_ON_OWNER
273 * Optimistic spinning.
275 * We try to spin for acquisition when we find that there are no
276 * pending waiters and the lock owner is currently running on a
279 * The rationale is that if the lock owner is running, it is likely to
280 * release the lock soon.
282 * Since this needs the lock owner, and this mutex implementation
283 * doesn't track the owner atomically in the lock field, we need to
284 * track it non-atomically.
286 * We can't do this for DEBUG_MUTEXES because that relies on wait_lock
287 * to serialize everything.
289 * The mutex spinners are queued up using MCS lock so that only one
290 * spinner can compete for the mutex. However, if mutex spinning isn't
291 * going to happen, there is no point in going through the lock/unlock
294 if (!mutex_can_spin_on_owner(lock))
298 struct task_struct *owner;
299 struct mspin_node node;
302 * If there's an owner, wait for it to either
303 * release the lock or go to sleep.
305 mspin_lock(MLOCK(lock), &node);
306 owner = ACCESS_ONCE(lock->owner);
307 if (owner && !mutex_spin_on_owner(lock, owner)) {
308 mspin_unlock(MLOCK(lock), &node);
312 if ((atomic_read(&lock->count) == 1) &&
313 (atomic_cmpxchg(&lock->count, 1, 0) == 1)) {
314 lock_acquired(&lock->dep_map, ip);
315 mutex_set_owner(lock);
316 mspin_unlock(MLOCK(lock), &node);
320 mspin_unlock(MLOCK(lock), &node);
323 * When there's no owner, we might have preempted between the
324 * owner acquiring the lock and setting the owner field. If
325 * we're an RT task that will live-lock because we won't let
326 * the owner complete.
328 if (!owner && (need_resched() || rt_task(task)))
332 * The cpu_relax() call is a compiler barrier which forces
333 * everything in this loop to be re-loaded. We don't need
334 * memory barriers as we'll eventually observe the right
335 * values at the cost of a few extra spins.
337 arch_mutex_cpu_relax();
341 spin_lock_mutex(&lock->wait_lock, flags);
343 debug_mutex_lock_common(lock, &waiter);
344 debug_mutex_add_waiter(lock, &waiter, task_thread_info(task));
346 /* add waiting tasks to the end of the waitqueue (FIFO): */
347 list_add_tail(&waiter.list, &lock->wait_list);
350 if (MUTEX_SHOW_NO_WAITER(lock) && (atomic_xchg(&lock->count, -1) == 1))
353 lock_contended(&lock->dep_map, ip);
357 * Lets try to take the lock again - this is needed even if
358 * we get here for the first time (shortly after failing to
359 * acquire the lock), to make sure that we get a wakeup once
360 * it's unlocked. Later on, if we sleep, this is the
361 * operation that gives us the lock. We xchg it to -1, so
362 * that when we release the lock, we properly wake up the
365 if (MUTEX_SHOW_NO_WAITER(lock) &&
366 (atomic_xchg(&lock->count, -1) == 1))
370 * got a signal? (This code gets eliminated in the
371 * TASK_UNINTERRUPTIBLE case.)
373 if (unlikely(signal_pending_state(state, task))) {
374 mutex_remove_waiter(lock, &waiter,
375 task_thread_info(task));
376 mutex_release(&lock->dep_map, 1, ip);
377 spin_unlock_mutex(&lock->wait_lock, flags);
379 debug_mutex_free_waiter(&waiter);
383 __set_task_state(task, state);
385 /* didn't get the lock, go to sleep: */
386 spin_unlock_mutex(&lock->wait_lock, flags);
387 schedule_preempt_disabled();
388 spin_lock_mutex(&lock->wait_lock, flags);
392 lock_acquired(&lock->dep_map, ip);
393 /* got the lock - rejoice! */
394 mutex_remove_waiter(lock, &waiter, current_thread_info());
395 mutex_set_owner(lock);
397 /* set it to 0 if there are no waiters left: */
398 if (likely(list_empty(&lock->wait_list)))
399 atomic_set(&lock->count, 0);
401 spin_unlock_mutex(&lock->wait_lock, flags);
403 debug_mutex_free_waiter(&waiter);
409 #ifdef CONFIG_DEBUG_LOCK_ALLOC
411 mutex_lock_nested(struct mutex *lock, unsigned int subclass)
414 __mutex_lock_common(lock, TASK_UNINTERRUPTIBLE, subclass, NULL, _RET_IP_);
417 EXPORT_SYMBOL_GPL(mutex_lock_nested);
420 _mutex_lock_nest_lock(struct mutex *lock, struct lockdep_map *nest)
423 __mutex_lock_common(lock, TASK_UNINTERRUPTIBLE, 0, nest, _RET_IP_);
426 EXPORT_SYMBOL_GPL(_mutex_lock_nest_lock);
429 mutex_lock_killable_nested(struct mutex *lock, unsigned int subclass)
432 return __mutex_lock_common(lock, TASK_KILLABLE, subclass, NULL, _RET_IP_);
434 EXPORT_SYMBOL_GPL(mutex_lock_killable_nested);
437 mutex_lock_interruptible_nested(struct mutex *lock, unsigned int subclass)
440 return __mutex_lock_common(lock, TASK_INTERRUPTIBLE,
441 subclass, NULL, _RET_IP_);
444 EXPORT_SYMBOL_GPL(mutex_lock_interruptible_nested);
448 * Release the lock, slowpath:
451 __mutex_unlock_common_slowpath(atomic_t *lock_count, int nested)
453 struct mutex *lock = container_of(lock_count, struct mutex, count);
456 spin_lock_mutex(&lock->wait_lock, flags);
457 mutex_release(&lock->dep_map, nested, _RET_IP_);
458 debug_mutex_unlock(lock);
461 * some architectures leave the lock unlocked in the fastpath failure
462 * case, others need to leave it locked. In the later case we have to
465 if (__mutex_slowpath_needs_to_unlock())
466 atomic_set(&lock->count, 1);
468 if (!list_empty(&lock->wait_list)) {
469 /* get the first entry from the wait-list: */
470 struct mutex_waiter *waiter =
471 list_entry(lock->wait_list.next,
472 struct mutex_waiter, list);
474 debug_mutex_wake_waiter(lock, waiter);
476 wake_up_process(waiter->task);
479 spin_unlock_mutex(&lock->wait_lock, flags);
483 * Release the lock, slowpath:
485 static __used noinline void
486 __mutex_unlock_slowpath(atomic_t *lock_count)
488 __mutex_unlock_common_slowpath(lock_count, 1);
491 #ifndef CONFIG_DEBUG_LOCK_ALLOC
493 * Here come the less common (and hence less performance-critical) APIs:
494 * mutex_lock_interruptible() and mutex_trylock().
496 static noinline int __sched
497 __mutex_lock_killable_slowpath(atomic_t *lock_count);
499 static noinline int __sched
500 __mutex_lock_interruptible_slowpath(atomic_t *lock_count);
503 * mutex_lock_interruptible - acquire the mutex, interruptible
504 * @lock: the mutex to be acquired
506 * Lock the mutex like mutex_lock(), and return 0 if the mutex has
507 * been acquired or sleep until the mutex becomes available. If a
508 * signal arrives while waiting for the lock then this function
511 * This function is similar to (but not equivalent to) down_interruptible().
513 int __sched mutex_lock_interruptible(struct mutex *lock)
518 ret = __mutex_fastpath_lock_retval
519 (&lock->count, __mutex_lock_interruptible_slowpath);
521 mutex_set_owner(lock);
526 EXPORT_SYMBOL(mutex_lock_interruptible);
528 int __sched mutex_lock_killable(struct mutex *lock)
533 ret = __mutex_fastpath_lock_retval
534 (&lock->count, __mutex_lock_killable_slowpath);
536 mutex_set_owner(lock);
540 EXPORT_SYMBOL(mutex_lock_killable);
542 static __used noinline void __sched
543 __mutex_lock_slowpath(atomic_t *lock_count)
545 struct mutex *lock = container_of(lock_count, struct mutex, count);
547 __mutex_lock_common(lock, TASK_UNINTERRUPTIBLE, 0, NULL, _RET_IP_);
550 static noinline int __sched
551 __mutex_lock_killable_slowpath(atomic_t *lock_count)
553 struct mutex *lock = container_of(lock_count, struct mutex, count);
555 return __mutex_lock_common(lock, TASK_KILLABLE, 0, NULL, _RET_IP_);
558 static noinline int __sched
559 __mutex_lock_interruptible_slowpath(atomic_t *lock_count)
561 struct mutex *lock = container_of(lock_count, struct mutex, count);
563 return __mutex_lock_common(lock, TASK_INTERRUPTIBLE, 0, NULL, _RET_IP_);
568 * Spinlock based trylock, we take the spinlock and check whether we
571 static inline int __mutex_trylock_slowpath(atomic_t *lock_count)
573 struct mutex *lock = container_of(lock_count, struct mutex, count);
577 spin_lock_mutex(&lock->wait_lock, flags);
579 prev = atomic_xchg(&lock->count, -1);
580 if (likely(prev == 1)) {
581 mutex_set_owner(lock);
582 mutex_acquire(&lock->dep_map, 0, 1, _RET_IP_);
585 /* Set it back to 0 if there are no waiters: */
586 if (likely(list_empty(&lock->wait_list)))
587 atomic_set(&lock->count, 0);
589 spin_unlock_mutex(&lock->wait_lock, flags);
595 * mutex_trylock - try to acquire the mutex, without waiting
596 * @lock: the mutex to be acquired
598 * Try to acquire the mutex atomically. Returns 1 if the mutex
599 * has been acquired successfully, and 0 on contention.
601 * NOTE: this function follows the spin_trylock() convention, so
602 * it is negated from the down_trylock() return values! Be careful
603 * about this when converting semaphore users to mutexes.
605 * This function must not be used in interrupt context. The
606 * mutex must be released by the same task that acquired it.
608 int __sched mutex_trylock(struct mutex *lock)
612 ret = __mutex_fastpath_trylock(&lock->count, __mutex_trylock_slowpath);
614 mutex_set_owner(lock);
618 EXPORT_SYMBOL(mutex_trylock);
621 * atomic_dec_and_mutex_lock - return holding mutex if we dec to 0
622 * @cnt: the atomic which we are to dec
623 * @lock: the mutex to return holding if we dec to 0
625 * return true and hold lock if we dec to 0, return false otherwise
627 int atomic_dec_and_mutex_lock(atomic_t *cnt, struct mutex *lock)
629 /* dec if we can't possibly hit 0 */
630 if (atomic_add_unless(cnt, -1, 1))
632 /* we might hit 0, so take the lock */
634 if (!atomic_dec_and_test(cnt)) {
635 /* when we actually did the dec, we didn't hit 0 */
639 /* we hit 0, and we hold the lock */
642 EXPORT_SYMBOL(atomic_dec_and_mutex_lock);