lock.h   [plain text]

/*
 * Copyright (c) 2016 Apple Inc. All rights reserved.
 *
 * @APPLE_APACHE_LICENSE_HEADER_START@
 *
 * Licensed under the Apache License, Version 2.0 (the "License");
 * you may not use this file except in compliance with the License.
 * You may obtain a copy of the License at
 *
 *     http://www.apache.org/licenses/LICENSE-2.0
 *
 * Unless required by applicable law or agreed to in writing, software
 * distributed under the License is distributed on an "AS IS" BASIS,
 * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
 * See the License for the specific language governing permissions and
 * limitations under the License.
 *
 * @APPLE_APACHE_LICENSE_HEADER_END@
 */

/*
 * IMPORTANT: This header file describes INTERNAL interfaces to libdispatch
 * which are subject to change in future releases of Mac OS X. Any applications
 * relying on these interfaces WILL break.
 */

#ifndef __DISPATCH_SHIMS_LOCK__
#define __DISPATCH_SHIMS_LOCK__

#pragma mark - platform macros

DISPATCH_ENUM(dispatch_lock_options, uint32_t,
		DLOCK_LOCK_NONE 			= 0x00000000,
		DLOCK_LOCK_DATA_CONTENTION  = 0x00010000,
);

#if TARGET_OS_MAC

typedef mach_port_t dispatch_lock_owner;
typedef uint32_t dispatch_lock;

#define DLOCK_OWNER_NULL			((dispatch_lock_owner)MACH_PORT_NULL)
#define DLOCK_OWNER_MASK			((dispatch_lock)0xfffffffc)
#define DLOCK_NOWAITERS_BIT			((dispatch_lock)0x00000001)
#define DLOCK_NOFAILED_TRYLOCK_BIT	((dispatch_lock)0x00000002)
#define _dispatch_tid_self()		((dispatch_lock_owner)_dispatch_thread_port())

DISPATCH_ALWAYS_INLINE
static inline bool
_dispatch_lock_is_locked(dispatch_lock lock_value)
{
	return (lock_value & DLOCK_OWNER_MASK) != 0;
}

DISPATCH_ALWAYS_INLINE
static inline dispatch_lock_owner
_dispatch_lock_owner(dispatch_lock lock_value)
{
	lock_value &= DLOCK_OWNER_MASK;
	if (lock_value) {
		lock_value |= DLOCK_NOWAITERS_BIT | DLOCK_NOFAILED_TRYLOCK_BIT;
	}
	return lock_value;
}

DISPATCH_ALWAYS_INLINE
static inline bool
_dispatch_lock_is_locked_by(dispatch_lock lock_value, dispatch_lock_owner tid)
{
	// equivalent to _dispatch_lock_owner(lock_value) == tid
	return ((lock_value ^ tid) & DLOCK_OWNER_MASK) == 0;
}

DISPATCH_ALWAYS_INLINE
static inline bool
_dispatch_lock_has_waiters(dispatch_lock lock_value)
{
	bool nowaiters_bit = (lock_value & DLOCK_NOWAITERS_BIT);
	return _dispatch_lock_is_locked(lock_value) != nowaiters_bit;
}

DISPATCH_ALWAYS_INLINE
static inline bool
_dispatch_lock_has_failed_trylock(dispatch_lock lock_value)
{
	return !(lock_value & DLOCK_NOFAILED_TRYLOCK_BIT);
}

#elif defined(__linux__)
#include <linux/futex.h>
#include <unistd.h>
#include <sys/syscall.h>   /* For SYS_xxx definitions */

typedef uint32_t dispatch_lock;
typedef pid_t dispatch_lock_owner;

#define DLOCK_OWNER_NULL			((dispatch_lock_owner)0)
#define DLOCK_OWNER_MASK			((dispatch_lock)FUTEX_TID_MASK)
#define DLOCK_WAITERS_BIT			((dispatch_lock)FUTEX_WAITERS)
#define DLOCK_FAILED_TRYLOCK_BIT	((dispatch_lock)FUTEX_OWNER_DIED)
#define _dispatch_tid_self() \
		((dispatch_lock_owner)(_dispatch_get_tsd_base()->tid))

DISPATCH_ALWAYS_INLINE
static inline bool
_dispatch_lock_is_locked(dispatch_lock lock_value)
{
	return (lock_value & DLOCK_OWNER_MASK) != 0;
}

DISPATCH_ALWAYS_INLINE
static inline dispatch_lock_owner
_dispatch_lock_owner(dispatch_lock lock_value)
{
	return (lock_value & DLOCK_OWNER_MASK);
}

DISPATCH_ALWAYS_INLINE
static inline bool
_dispatch_lock_is_locked_by(dispatch_lock lock_value, dispatch_lock_owner tid)
{
	return _dispatch_lock_owner(lock_value) == tid;
}

DISPATCH_ALWAYS_INLINE
static inline bool
_dispatch_lock_has_waiters(dispatch_lock lock_value)
{
	return (lock_value & DLOCK_WAITERS_BIT);
}

DISPATCH_ALWAYS_INLINE
static inline bool
_dispatch_lock_has_failed_trylock(dispatch_lock lock_value)
{
	return !(lock_value & DLOCK_FAILED_TRYLOCK_BIT);
}

#else
#  error define _dispatch_lock encoding scheme for your platform here
#endif

#if __has_include(<sys/ulock.h>)
#include <sys/ulock.h>
#endif

#ifndef HAVE_UL_COMPARE_AND_WAIT
#if defined(UL_COMPARE_AND_WAIT) && DISPATCH_HOST_SUPPORTS_OSX(101200)
#  define HAVE_UL_COMPARE_AND_WAIT 1
#else
#  define HAVE_UL_COMPARE_AND_WAIT 0
#endif
#endif // HAVE_UL_COMPARE_AND_WAIT

#ifndef HAVE_UL_UNFAIR_LOCK
#if defined(UL_UNFAIR_LOCK) && DISPATCH_HOST_SUPPORTS_OSX(101200)
#  define HAVE_UL_UNFAIR_LOCK 1
#else
#  define HAVE_UL_UNFAIR_LOCK 0
#endif
#endif // HAVE_UL_UNFAIR_LOCK

#ifndef DISPATCH_LOCK_USE_SEMAPHORE_FALLBACK
#define DISPATCH_LOCK_USE_SEMAPHORE_FALLBACK (!HAVE_UL_COMPARE_AND_WAIT && !HAVE_FUTEX)
#endif

#ifndef HAVE_FUTEX
#ifdef __linux__
#define HAVE_FUTEX 1
#else
#define HAVE_FUTEX 0
#endif
#endif // HAVE_FUTEX

#if USE_MACH_SEM
#define DISPATCH_SEMAPHORE_VERIFY_KR(x) do { \
		if (unlikely((x) == KERN_INVALID_NAME)) { \
			DISPATCH_CLIENT_CRASH((x), "Use-after-free of dispatch_semaphore_t"); \
		} else if (unlikely(x)) { \
			DISPATCH_INTERNAL_CRASH((x), "mach semaphore API failure"); \
		} \
	} while (0)
#define DISPATCH_GROUP_VERIFY_KR(x) do { \
		if (unlikely((x) == KERN_INVALID_NAME)) { \
			DISPATCH_CLIENT_CRASH((x), "Use-after-free of dispatch_group_t"); \
		} else if (unlikely(x)) { \
			DISPATCH_INTERNAL_CRASH((x), "mach semaphore API failure"); \
		} \
	} while (0)
#elif USE_POSIX_SEM
#define DISPATCH_SEMAPHORE_VERIFY_RET(x) do { \
		if (unlikely((x) == -1)) { \
			DISPATCH_INTERNAL_CRASH(errno, "POSIX semaphore API failure"); \
		} \
	} while (0)
#endif

#pragma mark - compare and wait

DISPATCH_NOT_TAIL_CALLED
void _dispatch_wait_on_address(uint32_t volatile *address, uint32_t value,
		dispatch_lock_options_t flags);
void _dispatch_wake_by_address(uint32_t volatile *address);

#pragma mark - thread event
/**
 * @typedef dispatch_thread_event_t
 *
 * @abstract
 * Dispatch Thread Events are used for one-time synchronization between threads.
 *
 * @discussion
 * Dispatch Thread Events are cheap synchronization points used when a thread
 * needs to block until a certain event has happened. Dispatch Thread Event
 * must be initialized and destroyed with _dispatch_thread_event_init() and
 * _dispatch_thread_event_destroy().
 *
 * A Dispatch Thread Event must be waited on and signaled exactly once between
 * initialization and destruction. These objects are simpler than semaphores
 * and do not support being signaled and waited on an arbitrary number of times.
 *
 * This locking primitive has no notion of ownership
 */
typedef struct dispatch_thread_event_s {
#if DISPATCH_LOCK_USE_SEMAPHORE_FALLBACK
	union {
		semaphore_t dte_semaphore;
		uint32_t dte_value;
	};
#elif HAVE_UL_COMPARE_AND_WAIT || HAVE_FUTEX
	// 1 means signalled but not waited on yet
	// UINT32_MAX means waited on, but not signalled yet
	// 0 is the initial and final state
	uint32_t dte_value;
#elif USE_POSIX_SEM
	sem_t dte_sem;
#else
#  error define dispatch_thread_event_s for your platform
#endif
} dispatch_thread_event_s, *dispatch_thread_event_t;

#if DISPATCH_LOCK_USE_SEMAPHORE_FALLBACK
semaphore_t _dispatch_thread_semaphore_create(void);
void _dispatch_thread_semaphore_dispose(void *);

DISPATCH_ALWAYS_INLINE
static inline semaphore_t
_dispatch_get_thread_semaphore(void)
{
	semaphore_t sema = (semaphore_t)(uintptr_t)
			_dispatch_thread_getspecific(dispatch_sema4_key);
	if (unlikely(!sema)) {
		return _dispatch_thread_semaphore_create();
	}
	_dispatch_thread_setspecific(dispatch_sema4_key, NULL);
	return sema;
}

DISPATCH_ALWAYS_INLINE
static inline void
_dispatch_put_thread_semaphore(semaphore_t sema)
{
	semaphore_t old_sema = (semaphore_t)(uintptr_t)
			_dispatch_thread_getspecific(dispatch_sema4_key);
	_dispatch_thread_setspecific(dispatch_sema4_key, (void*)(uintptr_t)sema);
	if (unlikely(old_sema)) {
		return _dispatch_thread_semaphore_dispose((void *)(uintptr_t)old_sema);
	}
}
#endif

DISPATCH_NOT_TAIL_CALLED
void _dispatch_thread_event_wait_slow(dispatch_thread_event_t);
void _dispatch_thread_event_signal_slow(dispatch_thread_event_t);

DISPATCH_ALWAYS_INLINE
static inline void
_dispatch_thread_event_init(dispatch_thread_event_t dte)
{
#if DISPATCH_LOCK_USE_SEMAPHORE_FALLBACK
	if (DISPATCH_LOCK_USE_SEMAPHORE_FALLBACK) {
		dte->dte_semaphore = _dispatch_get_thread_semaphore();
		return;
	}
#endif
#if HAVE_UL_COMPARE_AND_WAIT || HAVE_FUTEX
	dte->dte_value = 0;
#elif USE_POSIX_SEM
	int rc = sem_init(&dte->dte_sem, 0, 0);
	DISPATCH_SEMAPHORE_VERIFY_RET(rc);
#endif
}

DISPATCH_ALWAYS_INLINE
static inline void
_dispatch_thread_event_signal(dispatch_thread_event_t dte)
{
#if DISPATCH_LOCK_USE_SEMAPHORE_FALLBACK
	if (DISPATCH_LOCK_USE_SEMAPHORE_FALLBACK) {
		_dispatch_thread_event_signal_slow(dte);
		return;
	}
#endif
#if HAVE_UL_COMPARE_AND_WAIT || HAVE_FUTEX
	if (os_atomic_inc_orig(&dte->dte_value, release) == 0) {
		// 0 -> 1 transition doesn't need a signal
		// force a wake even when the value is corrupt,
		// waiters do the validation
		return;
	}
#elif USE_POSIX_SEM
	// fallthrough
#endif
	_dispatch_thread_event_signal_slow(dte);
}


DISPATCH_ALWAYS_INLINE
static inline void
_dispatch_thread_event_wait(dispatch_thread_event_t dte)
{
#if DISPATCH_LOCK_USE_SEMAPHORE_FALLBACK
	if (DISPATCH_LOCK_USE_SEMAPHORE_FALLBACK) {
		_dispatch_thread_event_wait_slow(dte);
		return;
	}
#endif
#if HAVE_UL_COMPARE_AND_WAIT || HAVE_FUTEX
	if (os_atomic_dec(&dte->dte_value, acquire) == 0) {
		// 1 -> 0 is always a valid transition, so we can return
		// for any other value, go to the slowpath which checks it's not corrupt
		return;
	}
#elif USE_POSIX_SEM
	// fallthrough
#endif
	_dispatch_thread_event_wait_slow(dte);
}

DISPATCH_ALWAYS_INLINE
static inline void
_dispatch_thread_event_destroy(dispatch_thread_event_t dte)
{
#if DISPATCH_LOCK_USE_SEMAPHORE_FALLBACK
	if (DISPATCH_LOCK_USE_SEMAPHORE_FALLBACK) {
		_dispatch_put_thread_semaphore(dte->dte_semaphore);
		return;
	}
#endif
#if HAVE_UL_COMPARE_AND_WAIT || HAVE_FUTEX
	// nothing to do
	dispatch_assert(dte->dte_value == 0);
#elif USE_POSIX_SEM
	int rc = sem_destroy(&dte->dte_sem);
	DISPATCH_SEMAPHORE_VERIFY_RET(rc);
#endif
}

#pragma mark - unfair lock

typedef struct dispatch_unfair_lock_s {
	dispatch_lock dul_lock;
} dispatch_unfair_lock_s, *dispatch_unfair_lock_t;

DISPATCH_NOT_TAIL_CALLED
void _dispatch_unfair_lock_lock_slow(dispatch_unfair_lock_t l,
		dispatch_lock_options_t options);
void _dispatch_unfair_lock_unlock_slow(dispatch_unfair_lock_t l,
		dispatch_lock tid_cur);

DISPATCH_ALWAYS_INLINE
static inline void
_dispatch_unfair_lock_lock(dispatch_unfair_lock_t l)
{
	dispatch_lock tid_self = _dispatch_tid_self();
	if (likely(os_atomic_cmpxchg(&l->dul_lock,
			DLOCK_OWNER_NULL, tid_self, acquire))) {
		return;
	}
	return _dispatch_unfair_lock_lock_slow(l, DLOCK_LOCK_NONE);
}

DISPATCH_ALWAYS_INLINE
static inline bool
_dispatch_unfair_lock_trylock(dispatch_unfair_lock_t l,
		dispatch_lock_owner *owner)
{
	dispatch_lock tid_old, tid_new, tid_self = _dispatch_tid_self();

	os_atomic_rmw_loop(&l->dul_lock, tid_old, tid_new, acquire, {
		if (likely(!_dispatch_lock_is_locked(tid_old))) {
			tid_new = tid_self;
		} else {
#ifdef DLOCK_NOFAILED_TRYLOCK_BIT
			tid_new = tid_old & ~DLOCK_NOFAILED_TRYLOCK_BIT;
#else
			tid_new = tid_old | DLOCK_FAILED_TRYLOCK_BIT;
#endif
		}
	});
	if (owner) *owner = _dispatch_lock_owner(tid_new);
	return !_dispatch_lock_is_locked(tid_old);
}

DISPATCH_ALWAYS_INLINE
static inline bool
_dispatch_unfair_lock_tryunlock(dispatch_unfair_lock_t l)
{
	dispatch_lock tid_old, tid_new;

	os_atomic_rmw_loop(&l->dul_lock, tid_old, tid_new, release, {
#ifdef DLOCK_NOFAILED_TRYLOCK_BIT
		if (likely(tid_old & DLOCK_NOFAILED_TRYLOCK_BIT)) {
			tid_new = DLOCK_OWNER_NULL;
		} else {
			tid_new = tid_old | DLOCK_NOFAILED_TRYLOCK_BIT;
		}
#else
		if (likely(!(tid_old & DLOCK_FAILED_TRYLOCK_BIT))) {
			tid_new = DLOCK_OWNER_NULL;
		} else {
			tid_new = tid_old & ~DLOCK_FAILED_TRYLOCK_BIT;
		}
#endif
	});
	if (unlikely(tid_new)) {
		// unlock failed, renew the lock, which needs an acquire barrier
		os_atomic_thread_fence(acquire);
		return false;
	}
	if (unlikely(_dispatch_lock_has_waiters(tid_old))) {
		_dispatch_unfair_lock_unlock_slow(l, tid_old);
	}
	return true;
}

DISPATCH_ALWAYS_INLINE
static inline bool
_dispatch_unfair_lock_unlock_had_failed_trylock(dispatch_unfair_lock_t l)
{
	dispatch_lock tid_cur, tid_self = _dispatch_tid_self();
#if HAVE_FUTEX
	if (likely(os_atomic_cmpxchgv(&l->dul_lock,
			tid_self, DLOCK_OWNER_NULL, &tid_cur, release))) {
		return false;
	}
#else
	tid_cur = os_atomic_xchg(&l->dul_lock, DLOCK_OWNER_NULL, release);
	if (likely(tid_cur == tid_self)) return false;
#endif
	_dispatch_unfair_lock_unlock_slow(l, tid_cur);
	return _dispatch_lock_has_failed_trylock(tid_cur);
}

DISPATCH_ALWAYS_INLINE
static inline void
_dispatch_unfair_lock_unlock(dispatch_unfair_lock_t l)
{
	(void)_dispatch_unfair_lock_unlock_had_failed_trylock(l);
}

#pragma mark - gate lock

#if HAVE_UL_UNFAIR_LOCK || HAVE_FUTEX
#define DISPATCH_GATE_USE_FOR_DISPATCH_ONCE 1
#else
#define DISPATCH_GATE_USE_FOR_DISPATCH_ONCE 0
#endif

#define DLOCK_GATE_UNLOCKED	((dispatch_lock)0)

#define DLOCK_ONCE_UNLOCKED	((dispatch_once_t)0)
#define DLOCK_ONCE_DONE		(~(dispatch_once_t)0)

typedef struct dispatch_gate_s {
	dispatch_lock dgl_lock;
} dispatch_gate_s, *dispatch_gate_t;

typedef struct dispatch_once_gate_s {
	union {
		dispatch_gate_s dgo_gate;
		dispatch_once_t dgo_once;
	};
} dispatch_once_gate_s, *dispatch_once_gate_t;

DISPATCH_NOT_TAIL_CALLED
void _dispatch_gate_wait_slow(dispatch_gate_t l, dispatch_lock value,
		uint32_t flags);
void _dispatch_gate_broadcast_slow(dispatch_gate_t l, dispatch_lock tid_cur);

DISPATCH_ALWAYS_INLINE
static inline bool
_dispatch_gate_tryenter(dispatch_gate_t l)
{
	dispatch_lock tid_self = _dispatch_tid_self();
	return likely(os_atomic_cmpxchg(&l->dgl_lock,
			DLOCK_GATE_UNLOCKED, tid_self, acquire));
}

#define _dispatch_gate_wait(l, flags) \
	_dispatch_gate_wait_slow(l, DLOCK_GATE_UNLOCKED, flags)

DISPATCH_ALWAYS_INLINE
static inline void
_dispatch_gate_broadcast(dispatch_gate_t l)
{
	dispatch_lock tid_cur, tid_self = _dispatch_tid_self();
	tid_cur = os_atomic_xchg(&l->dgl_lock, DLOCK_GATE_UNLOCKED, release);
	if (likely(tid_cur == tid_self)) return;
	_dispatch_gate_broadcast_slow(l, tid_cur);
}

DISPATCH_ALWAYS_INLINE
static inline bool
_dispatch_once_gate_tryenter(dispatch_once_gate_t l)
{
	dispatch_once_t tid_self = (dispatch_once_t)_dispatch_tid_self();
	return likely(os_atomic_cmpxchg(&l->dgo_once,
			DLOCK_ONCE_UNLOCKED, tid_self, acquire));
}

#define _dispatch_once_gate_wait(l) \
	_dispatch_gate_wait_slow(&(l)->dgo_gate, (dispatch_lock)DLOCK_ONCE_DONE, \
			DLOCK_LOCK_NONE)

DISPATCH_ALWAYS_INLINE
static inline void
_dispatch_once_gate_broadcast(dispatch_once_gate_t l)
{
	dispatch_once_t tid_cur, tid_self = (dispatch_once_t)_dispatch_tid_self();
	// see once.c for explanation about this trick
	os_atomic_maximally_synchronizing_barrier();
	// above assumed to contain release barrier
	tid_cur = os_atomic_xchg(&l->dgo_once, DLOCK_ONCE_DONE, relaxed);
	if (likely(tid_cur == tid_self)) return;
	_dispatch_gate_broadcast_slow(&l->dgo_gate, (dispatch_lock)tid_cur);
}

#endif // __DISPATCH_SHIMS_LOCK__