[net-next-2.6.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>

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