[linux-block.git] / lib / rwsem.c

/* rwsem.c: R/W semaphores: contention handling functions
 *
 * Written by David Howells (dhowells@redhat.com).
 * Derived from arch/i386/kernel/semaphore.c
 */
#include <linux/rwsem.h>
#include <linux/sched.h>
#include <linux/init.h>
#include <linux/module.h>

/*
 * Initialize an rwsem:
 */
void __init_rwsem(struct rw_semaphore *sem, const char *name,
		  struct lock_class_key *key)
{
#ifdef CONFIG_DEBUG_LOCK_ALLOC
	/*
	 * Make sure we are not reinitializing a held semaphore:
	 */
	debug_check_no_locks_freed((void *)sem, sizeof(*sem));
	lockdep_init_map(&sem->dep_map, name, key, 0);
#endif
	sem->count = RWSEM_UNLOCKED_VALUE;
	spin_lock_init(&sem->wait_lock);
	INIT_LIST_HEAD(&sem->wait_list);
}

EXPORT_SYMBOL(__init_rwsem);

struct rwsem_waiter {
	struct list_head list;
	struct task_struct *task;
	unsigned int flags;
#define RWSEM_WAITING_FOR_READ	0x00000001
#define RWSEM_WAITING_FOR_WRITE	0x00000002
};

/*
 * handle the lock release when processes blocked on it that can now run
 * - if we come here from up_xxxx(), then:
 *   - the 'active part' of count (&0x0000ffff) reached 0 (but may have changed)
 *   - the 'waiting part' of count (&0xffff0000) is -ve (and will still be so)
 *   - there must be someone on the queue
 * - the spinlock must be held by the caller
 * - woken process blocks are discarded from the list after having task zeroed
 * - writers are only woken if downgrading is false
 */
static inline struct rw_semaphore *
__rwsem_do_wake(struct rw_semaphore *sem, int downgrading)
{
	struct rwsem_waiter *waiter;
	struct task_struct *tsk;
	struct list_head *next;
	signed long oldcount, woken, loop;

	if (downgrading)
		goto dont_wake_writers;

	/* if we came through an up_xxxx() call, we only only wake someone up
	 * if we can transition the active part of the count from 0 -> 1
	 */
 try_again:
	oldcount = rwsem_atomic_update(RWSEM_ACTIVE_BIAS, sem)
						- RWSEM_ACTIVE_BIAS;
	if (oldcount & RWSEM_ACTIVE_MASK)
		goto undo;

	waiter = list_entry(sem->wait_list.next, struct rwsem_waiter, list);

	/* try to grant a single write lock if there's a writer at the front
	 * of the queue - note we leave the 'active part' of the count
	 * incremented by 1 and the waiting part incremented by 0x00010000
	 */
	if (!(waiter->flags & RWSEM_WAITING_FOR_WRITE))
		goto readers_only;

	/* We must be careful not to touch 'waiter' after we set ->task = NULL.
	 * It is an allocated on the waiter's stack and may become invalid at
	 * any time after that point (due to a wakeup from another source).
	 */
	list_del(&waiter->list);
	tsk = waiter->task;
	smp_mb();
	waiter->task = NULL;
	wake_up_process(tsk);
	put_task_struct(tsk);
	goto out;

	/* don't want to wake any writers */
 dont_wake_writers:
	waiter = list_entry(sem->wait_list.next, struct rwsem_waiter, list);
	if (waiter->flags & RWSEM_WAITING_FOR_WRITE)
		goto out;

	/* grant an infinite number of read locks to the readers at the front
	 * of the queue
	 * - note we increment the 'active part' of the count by the number of
	 *   readers before waking any processes up
	 */
 readers_only:
	woken = 0;
	do {
		woken++;

		if (waiter->list.next == &sem->wait_list)
			break;

		waiter = list_entry(waiter->list.next,
					struct rwsem_waiter, list);

	} while (waiter->flags & RWSEM_WAITING_FOR_READ);

	loop = woken;
	woken *= RWSEM_ACTIVE_BIAS - RWSEM_WAITING_BIAS;
	if (!downgrading)
		/* we'd already done one increment earlier */
		woken -= RWSEM_ACTIVE_BIAS;

	rwsem_atomic_add(woken, sem);

	next = sem->wait_list.next;
	for (; loop > 0; loop--) {
		waiter = list_entry(next, struct rwsem_waiter, list);
		next = waiter->list.next;
		tsk = waiter->task;
		smp_mb();
		waiter->task = NULL;
		wake_up_process(tsk);
		put_task_struct(tsk);
	}

	sem->wait_list.next = next;
	next->prev = &sem->wait_list;

 out:
	return sem;

	/* undo the change to count, but check for a transition 1->0 */
 undo:
	if (rwsem_atomic_update(-RWSEM_ACTIVE_BIAS, sem) != 0)
		goto out;
	goto try_again;
}

/*
 * wait for a lock to be granted
 */
static struct rw_semaphore __sched *
rwsem_down_failed_common(struct rw_semaphore *sem,
			struct rwsem_waiter *waiter, signed long adjustment)
{
	struct task_struct *tsk = current;
	signed long count;

	set_task_state(tsk, TASK_UNINTERRUPTIBLE);

	/* set up my own style of waitqueue */
	spin_lock_irq(&sem->wait_lock);
	waiter->task = tsk;
	get_task_struct(tsk);

	list_add_tail(&waiter->list, &sem->wait_list);

	/* we're now waiting on the lock, but no longer actively read-locking */
	count = rwsem_atomic_update(adjustment, sem);

	/* if there are no active locks, wake the front queued process(es) up */
	if (!(count & RWSEM_ACTIVE_MASK))
		sem = __rwsem_do_wake(sem, 0);

	spin_unlock_irq(&sem->wait_lock);

	/* wait to be given the lock */
	for (;;) {
		if (!waiter->task)
			break;
		schedule();
		set_task_state(tsk, TASK_UNINTERRUPTIBLE);
	}

	tsk->state = TASK_RUNNING;

	return sem;
}

/*
 * wait for the read lock to be granted
 */
asmregparm struct rw_semaphore __sched *
rwsem_down_read_failed(struct rw_semaphore *sem)
{
	struct rwsem_waiter waiter;

	waiter.flags = RWSEM_WAITING_FOR_READ;
	rwsem_down_failed_common(sem, &waiter,
				RWSEM_WAITING_BIAS - RWSEM_ACTIVE_BIAS);
	return sem;
}

/*
 * wait for the write lock to be granted
 */
asmregparm struct rw_semaphore __sched *
rwsem_down_write_failed(struct rw_semaphore *sem)
{
	struct rwsem_waiter waiter;

	waiter.flags = RWSEM_WAITING_FOR_WRITE;
	rwsem_down_failed_common(sem, &waiter, -RWSEM_ACTIVE_BIAS);

	return sem;
}

/*
 * handle waking up a waiter on the semaphore
 * - up_read/up_write has decremented the active part of count if we come here
 */
asmregparm struct rw_semaphore *rwsem_wake(struct rw_semaphore *sem)
{
	unsigned long flags;

	spin_lock_irqsave(&sem->wait_lock, flags);

	/* do nothing if list empty */
	if (!list_empty(&sem->wait_list))
		sem = __rwsem_do_wake(sem, 0);

	spin_unlock_irqrestore(&sem->wait_lock, flags);

	return sem;
}

/*
 * downgrade a write lock into a read lock
 * - caller incremented waiting part of count and discovered it still negative
 * - just wake up any readers at the front of the queue
 */
asmregparm struct rw_semaphore *rwsem_downgrade_wake(struct rw_semaphore *sem)
{
	unsigned long flags;

	spin_lock_irqsave(&sem->wait_lock, flags);

	/* do nothing if list empty */
	if (!list_empty(&sem->wait_list))
		sem = __rwsem_do_wake(sem, 1);

	spin_unlock_irqrestore(&sem->wait_lock, flags);

	return sem;
}

EXPORT_SYMBOL(rwsem_down_read_failed);
EXPORT_SYMBOL(rwsem_down_write_failed);
EXPORT_SYMBOL(rwsem_wake);
EXPORT_SYMBOL(rwsem_downgrade_wake);
Commit	Line	Data
1da177e4 LT	1	/* rwsem.c: R/W semaphores: contention handling functions
	2	*
	3	* Written by David Howells (dhowells@redhat.com).
	4	* Derived from arch/i386/kernel/semaphore.c
	5	*/
	6	#include <linux/rwsem.h>
	7	#include <linux/sched.h>
	8	#include <linux/init.h>
	9	#include <linux/module.h>
	10
4ea2176d IM	11	/*
	12	* Initialize an rwsem:
	13	*/
	14	void __init_rwsem(struct rw_semaphore sem, const char name,
	15	struct lock_class_key *key)
	16	{
	17	#ifdef CONFIG_DEBUG_LOCK_ALLOC
	18	/*
	19	* Make sure we are not reinitializing a held semaphore:
	20	*/
	21	debug_check_no_locks_freed((void )sem, sizeof(sem));
4dfbb9d8	22	lockdep_init_map(&sem->dep_map, name, key, 0);
4ea2176d IM	23	#endif
	24	sem->count = RWSEM_UNLOCKED_VALUE;
	25	spin_lock_init(&sem->wait_lock);
	26	INIT_LIST_HEAD(&sem->wait_list);
	27	}
	28
	29	EXPORT_SYMBOL(__init_rwsem);
	30
1da177e4 LT	31	struct rwsem_waiter {
	32	struct list_head list;
	33	struct task_struct *task;
	34	unsigned int flags;
	35	#define RWSEM_WAITING_FOR_READ 0x00000001
	36	#define RWSEM_WAITING_FOR_WRITE 0x00000002
	37	};
	38
1da177e4 LT	39	/*
	40	* handle the lock release when processes blocked on it that can now run
	41	* - if we come here from up_xxxx(), then:
	42	* - the 'active part' of count (&0x0000ffff) reached 0 (but may have changed)
	43	* - the 'waiting part' of count (&0xffff0000) is -ve (and will still be so)
	44	* - there must be someone on the queue
	45	* - the spinlock must be held by the caller
	46	* - woken process blocks are discarded from the list after having task zeroed
	47	* - writers are only woken if downgrading is false
	48	*/
	49	static inline struct rw_semaphore *
	50	__rwsem_do_wake(struct rw_semaphore *sem, int downgrading)
	51	{
	52	struct rwsem_waiter *waiter;
	53	struct task_struct *tsk;
	54	struct list_head *next;
	55	signed long oldcount, woken, loop;
	56
1da177e4 LT	57	if (downgrading)
	58	goto dont_wake_writers;
	59
	60	/* if we came through an up_xxxx() call, we only only wake someone up
	61	* if we can transition the active part of the count from 0 -> 1
	62	*/
	63	try_again:
	64	oldcount = rwsem_atomic_update(RWSEM_ACTIVE_BIAS, sem)
	65	- RWSEM_ACTIVE_BIAS;
	66	if (oldcount & RWSEM_ACTIVE_MASK)
	67	goto undo;
	68
	69	waiter = list_entry(sem->wait_list.next, struct rwsem_waiter, list);
	70
	71	/* try to grant a single write lock if there's a writer at the front
	72	* of the queue - note we leave the 'active part' of the count
	73	* incremented by 1 and the waiting part incremented by 0x00010000
	74	*/
	75	if (!(waiter->flags & RWSEM_WAITING_FOR_WRITE))
	76	goto readers_only;
	77
	78	/* We must be careful not to touch 'waiter' after we set ->task = NULL.
	79	* It is an allocated on the waiter's stack and may become invalid at
	80	* any time after that point (due to a wakeup from another source).
	81	*/
	82	list_del(&waiter->list);
	83	tsk = waiter->task;
d59dd462	84	smp_mb();
1da177e4 LT	85	waiter->task = NULL;
	86	wake_up_process(tsk);
	87	put_task_struct(tsk);
	88	goto out;
	89
	90	/* don't want to wake any writers */
	91	dont_wake_writers:
	92	waiter = list_entry(sem->wait_list.next, struct rwsem_waiter, list);
	93	if (waiter->flags & RWSEM_WAITING_FOR_WRITE)
	94	goto out;
	95
	96	/* grant an infinite number of read locks to the readers at the front
	97	* of the queue
	98	* - note we increment the 'active part' of the count by the number of
	99	* readers before waking any processes up
	100	*/
	101	readers_only:
	102	woken = 0;
	103	do {
	104	woken++;
	105
	106	if (waiter->list.next == &sem->wait_list)
	107	break;
	108
	109	waiter = list_entry(waiter->list.next,
	110	struct rwsem_waiter, list);
	111
	112	} while (waiter->flags & RWSEM_WAITING_FOR_READ);
	113
	114	loop = woken;
	115	woken *= RWSEM_ACTIVE_BIAS - RWSEM_WAITING_BIAS;
	116	if (!downgrading)
	117	/* we'd already done one increment earlier */
	118	woken -= RWSEM_ACTIVE_BIAS;
	119
	120	rwsem_atomic_add(woken, sem);
	121
	122	next = sem->wait_list.next;
	123	for (; loop > 0; loop--) {
	124	waiter = list_entry(next, struct rwsem_waiter, list);
	125	next = waiter->list.next;
	126	tsk = waiter->task;
d59dd462	127	smp_mb();
1da177e4 LT	128	waiter->task = NULL;
	129	wake_up_process(tsk);
	130	put_task_struct(tsk);
	131	}
	132
	133	sem->wait_list.next = next;
	134	next->prev = &sem->wait_list;
	135
	136	out:
1da177e4 LT	137	return sem;
	138
	139	/* undo the change to count, but check for a transition 1->0 */
	140	undo:
	141	if (rwsem_atomic_update(-RWSEM_ACTIVE_BIAS, sem) != 0)
	142	goto out;
	143	goto try_again;
	144	}
	145
	146	/*
	147	* wait for a lock to be granted
	148	*/
c7af77b5	149	static struct rw_semaphore __sched *
1da177e4 LT	150	rwsem_down_failed_common(struct rw_semaphore *sem,
	151	struct rwsem_waiter *waiter, signed long adjustment)
	152	{
	153	struct task_struct *tsk = current;
	154	signed long count;
	155
	156	set_task_state(tsk, TASK_UNINTERRUPTIBLE);
	157
	158	/* set up my own style of waitqueue */
	159	spin_lock_irq(&sem->wait_lock);
	160	waiter->task = tsk;
	161	get_task_struct(tsk);
	162
	163	list_add_tail(&waiter->list, &sem->wait_list);
	164
	165	/* we're now waiting on the lock, but no longer actively read-locking */
	166	count = rwsem_atomic_update(adjustment, sem);
	167
	168	/* if there are no active locks, wake the front queued process(es) up */
	169	if (!(count & RWSEM_ACTIVE_MASK))
	170	sem = __rwsem_do_wake(sem, 0);
	171
	172	spin_unlock_irq(&sem->wait_lock);
	173
	174	/* wait to be given the lock */
	175	for (;;) {
	176	if (!waiter->task)
	177	break;
	178	schedule();
	179	set_task_state(tsk, TASK_UNINTERRUPTIBLE);
	180	}
	181
	182	tsk->state = TASK_RUNNING;
	183
	184	return sem;
	185	}
	186
	187	/*
	188	* wait for the read lock to be granted
	189	*/
d50efc6c	190	asmregparm struct rw_semaphore __sched *
1da177e4 LT	191	rwsem_down_read_failed(struct rw_semaphore *sem)
	192	{
	193	struct rwsem_waiter waiter;
	194
1da177e4 LT	195	waiter.flags = RWSEM_WAITING_FOR_READ;
	196	rwsem_down_failed_common(sem, &waiter,
	197	RWSEM_WAITING_BIAS - RWSEM_ACTIVE_BIAS);
1da177e4 LT	198	return sem;
	199	}
	200
	201	/*
	202	* wait for the write lock to be granted
	203	*/
d50efc6c	204	asmregparm struct rw_semaphore __sched *
1da177e4 LT	205	rwsem_down_write_failed(struct rw_semaphore *sem)
	206	{
	207	struct rwsem_waiter waiter;
	208
1da177e4 LT	209	waiter.flags = RWSEM_WAITING_FOR_WRITE;
	210	rwsem_down_failed_common(sem, &waiter, -RWSEM_ACTIVE_BIAS);
	211
1da177e4 LT	212	return sem;
	213	}
	214
	215	/*
	216	* handle waking up a waiter on the semaphore
	217	* - up_read/up_write has decremented the active part of count if we come here
	218	*/
d50efc6c	219	asmregparm struct rw_semaphore rwsem_wake(struct rw_semaphore sem)
1da177e4 LT	220	{
	221	unsigned long flags;
	222
1da177e4 LT	223	spin_lock_irqsave(&sem->wait_lock, flags);
	224
	225	/* do nothing if list empty */
	226	if (!list_empty(&sem->wait_list))
	227	sem = __rwsem_do_wake(sem, 0);
	228
	229	spin_unlock_irqrestore(&sem->wait_lock, flags);
	230
1da177e4 LT	231	return sem;
	232	}
	233
	234	/*
	235	* downgrade a write lock into a read lock
	236	* - caller incremented waiting part of count and discovered it still negative
	237	* - just wake up any readers at the front of the queue
	238	*/
d50efc6c	239	asmregparm struct rw_semaphore rwsem_downgrade_wake(struct rw_semaphore sem)
1da177e4 LT	240	{
	241	unsigned long flags;
	242
1da177e4 LT	243	spin_lock_irqsave(&sem->wait_lock, flags);
	244
	245	/* do nothing if list empty */
	246	if (!list_empty(&sem->wait_list))
	247	sem = __rwsem_do_wake(sem, 1);
	248
	249	spin_unlock_irqrestore(&sem->wait_lock, flags);
	250
1da177e4 LT	251	return sem;
	252	}
	253
	254	EXPORT_SYMBOL(rwsem_down_read_failed);
	255	EXPORT_SYMBOL(rwsem_down_write_failed);
	256	EXPORT_SYMBOL(rwsem_wake);
	257	EXPORT_SYMBOL(rwsem_downgrade_wake);