dm.c

Linux Kernel 2.6.9 for OMAP1710

			return 0;	/* deferred successfully */

		/*
		 * We're in a while loop, because someone could suspend
		 * before we get to the following read lock.
		 */
		down_read(&md->lock);
	}

	__split_bio(md, bio);
	up_read(&md->lock);
	return 0;
}

static int dm_flush_all(request_queue_t *q, struct gendisk *disk,
			sector_t *error_sector)
{
	struct mapped_device *md = q->queuedata;
	struct dm_table *map = dm_get_table(md);
	int ret = -ENXIO;

	if (map) {
		ret = dm_table_flush_all(md->map);
		dm_table_put(map);
	}

	return ret;
}

static void dm_unplug_all(request_queue_t *q)
{
	struct mapped_device *md = q->queuedata;
	struct dm_table *map = dm_get_table(md);

	if (map) {
		dm_table_unplug_all(map);
		dm_table_put(map);
	}
}

static int dm_any_congested(void *congested_data, int bdi_bits)
{
	int r;
	struct mapped_device *md = (struct mapped_device *) congested_data;
	struct dm_table *map = dm_get_table(md);

	if (!map || test_bit(DMF_BLOCK_IO, &md->flags))
		r = bdi_bits;
	else
		r = dm_table_any_congested(map, bdi_bits);

	dm_table_put(map);
	return r;
}

/*-----------------------------------------------------------------
 * An IDR is used to keep track of allocated minor numbers.
 *---------------------------------------------------------------*/
static DECLARE_MUTEX(_minor_lock);
static DEFINE_IDR(_minor_idr);

static void free_minor(unsigned int minor)
{
	down(&_minor_lock);
	idr_remove(&_minor_idr, minor);
	up(&_minor_lock);
}

/*
 * See if the device with a specific minor # is free.
 */
static int specific_minor(unsigned int minor)
{
	int r, m;

	if (minor >= (1 << MINORBITS))
		return -EINVAL;

	down(&_minor_lock);

	if (idr_find(&_minor_idr, minor)) {
		r = -EBUSY;
		goto out;
	}

	r = idr_pre_get(&_minor_idr, GFP_KERNEL);
	if (!r) {
		r = -ENOMEM;
		goto out;
	}

	r = idr_get_new_above(&_minor_idr, specific_minor, minor, &m);
	if (r) {
		goto out;
	}

	if (m != minor) {
		idr_remove(&_minor_idr, m);
		r = -EBUSY;
		goto out;
	}

out:
	up(&_minor_lock);
	return r;
}

static int next_free_minor(unsigned int *minor)
{
	int r;
	unsigned int m;

	down(&_minor_lock);

	r = idr_pre_get(&_minor_idr, GFP_KERNEL);
	if (!r) {
		r = -ENOMEM;
		goto out;
	}

	r = idr_get_new(&_minor_idr, next_free_minor, &m);
	if (r) {
		goto out;
	}

	if (m >= (1 << MINORBITS)) {
		idr_remove(&_minor_idr, m);
		r = -ENOSPC;
		goto out;
	}

	*minor = m;

out:
	up(&_minor_lock);
	return r;
}

static struct block_device_operations dm_blk_dops;

/*
 * Allocate and initialise a blank device with a given minor.
 */
static struct mapped_device *alloc_dev(unsigned int minor, int persistent)
{
	int r;
	struct mapped_device *md = kmalloc(sizeof(*md), GFP_KERNEL);

	if (!md) {
		DMWARN("unable to allocate device, out of memory.");
		return NULL;
	}

	/* get a minor number for the dev */
	r = persistent ? specific_minor(minor) : next_free_minor(&minor);
	if (r < 0)
		goto bad1;

	memset(md, 0, sizeof(*md));
	init_rwsem(&md->lock);
	rwlock_init(&md->map_lock);
	atomic_set(&md->holders, 1);
	atomic_set(&md->event_nr, 0);

	md->queue = blk_alloc_queue(GFP_KERNEL);
	if (!md->queue)
		goto bad1;

	md->queue->queuedata = md;
	md->queue->backing_dev_info.congested_fn = dm_any_congested;
	md->queue->backing_dev_info.congested_data = md;
	blk_queue_make_request(md->queue, dm_request);
	md->queue->unplug_fn = dm_unplug_all;
	md->queue->issue_flush_fn = dm_flush_all;

	md->io_pool = mempool_create(MIN_IOS, mempool_alloc_slab,
				     mempool_free_slab, _io_cache);
 	if (!md->io_pool)
 		goto bad2;

	md->tio_pool = mempool_create(MIN_IOS, mempool_alloc_slab,
				      mempool_free_slab, _tio_cache);
	if (!md->tio_pool)
		goto bad3;

	md->disk = alloc_disk(1);
	if (!md->disk)
		goto bad4;

	md->disk->major = _major;
	md->disk->first_minor = minor;
	md->disk->fops = &dm_blk_dops;
	md->disk->queue = md->queue;
	md->disk->private_data = md;
	sprintf(md->disk->disk_name, "dm-%d", minor);
	add_disk(md->disk);

	atomic_set(&md->pending, 0);
	init_waitqueue_head(&md->wait);
	init_waitqueue_head(&md->eventq);

	return md;

 bad4:
	mempool_destroy(md->tio_pool);
 bad3:
	mempool_destroy(md->io_pool);
 bad2:
	blk_put_queue(md->queue);
	free_minor(minor);
 bad1:
	kfree(md);
	return NULL;
}

static void free_dev(struct mapped_device *md)
{
	free_minor(md->disk->first_minor);
	mempool_destroy(md->tio_pool);
	mempool_destroy(md->io_pool);
	del_gendisk(md->disk);
	put_disk(md->disk);
	blk_put_queue(md->queue);
	kfree(md);
}

/*
 * Bind a table to the device.
 */
static void event_callback(void *context)
{
	struct mapped_device *md = (struct mapped_device *) context;

	atomic_inc(&md->event_nr);;
	wake_up(&md->eventq);
}

static void __set_size(struct gendisk *disk, sector_t size)
{
	struct block_device *bdev;

	set_capacity(disk, size);
	bdev = bdget_disk(disk, 0);
	if (bdev) {
		down(&bdev->bd_inode->i_sem);
		i_size_write(bdev->bd_inode, (loff_t)size << SECTOR_SHIFT);
		up(&bdev->bd_inode->i_sem);
		bdput(bdev);
	}
}

static int __bind(struct mapped_device *md, struct dm_table *t)
{
	request_queue_t *q = md->queue;
	sector_t size;

	size = dm_table_get_size(t);
	__set_size(md->disk, size);
	if (size == 0)
		return 0;

	write_lock(&md->map_lock);
	md->map = t;
	write_unlock(&md->map_lock);

	dm_table_get(t);
	dm_table_event_callback(md->map, event_callback, md);
	dm_table_set_restrictions(t, q);
	return 0;
}

static void __unbind(struct mapped_device *md)
{
	struct dm_table *map = md->map;

	if (!map)
		return;

	dm_table_event_callback(map, NULL, NULL);
	write_lock(&md->map_lock);
	md->map = NULL;
	write_unlock(&md->map_lock);
	dm_table_put(map);
}

/*
 * Constructor for a new device.
 */
static int create_aux(unsigned int minor, int persistent,
		      struct mapped_device **result)
{
	struct mapped_device *md;

	md = alloc_dev(minor, persistent);
	if (!md)
		return -ENXIO;

	*result = md;
	return 0;
}

int dm_create(struct mapped_device **result)
{
	return create_aux(0, 0, result);
}

int dm_create_with_minor(unsigned int minor, struct mapped_device **result)
{
	return create_aux(minor, 1, result);
}

void dm_get(struct mapped_device *md)
{
	atomic_inc(&md->holders);
}

void dm_put(struct mapped_device *md)
{
	struct dm_table *map = dm_get_table(md);

	if (atomic_dec_and_test(&md->holders)) {
		if (!test_bit(DMF_SUSPENDED, &md->flags) && map)
			dm_table_suspend_targets(map);
		__unbind(md);
		free_dev(md);
	}

	dm_table_put(map);
}

/*
 * Process the deferred bios
 */
static void __flush_deferred_io(struct mapped_device *md, struct bio *c)
{
	struct bio *n;

	while (c) {
		n = c->bi_next;
		c->bi_next = NULL;
		__split_bio(md, c);
		c = n;
	}
}

/*
 * Swap in a new table (destroying old one).
 */
int dm_swap_table(struct mapped_device *md, struct dm_table *table)
{
	int r;

	down_write(&md->lock);

	/* device must be suspended */
	if (!test_bit(DMF_SUSPENDED, &md->flags)) {
		up_write(&md->lock);
		return -EPERM;
	}

	__unbind(md);
	r = __bind(md, table);
	if (r)
		return r;

	up_write(&md->lock);
	return 0;
}

/*
 * Functions to lock and unlock any filesystem running on the
 * device.
 */
static int __lock_fs(struct mapped_device *md)
{
	struct block_device *bdev;

	if (test_and_set_bit(DMF_FS_LOCKED, &md->flags))
		return 0;

	bdev = bdget_disk(md->disk, 0);
	if (!bdev) {
		DMWARN("bdget failed in __lock_fs");
		return -ENOMEM;
	}

	WARN_ON(md->frozen_sb);
	md->frozen_sb = freeze_bdev(bdev);
	/* don't bdput right now, we don't want the bdev
	 * to go away while it is locked.  We'll bdput
	 * in __unlock_fs
	 */
	return 0;
}

static int __unlock_fs(struct mapped_device *md)
{
	struct block_device *bdev;

	if (!test_and_clear_bit(DMF_FS_LOCKED, &md->flags))
		return 0;

	bdev = bdget_disk(md->disk, 0);
	if (!bdev) {
		DMWARN("bdget failed in __unlock_fs");
		return -ENOMEM;
	}

	thaw_bdev(bdev, md->frozen_sb);
	md->frozen_sb = NULL;
	bdput(bdev);
	bdput(bdev);
	return 0;
}

/*
 * We need to be able to change a mapping table under a mounted
 * filesystem.  For example we might want to move some data in
 * the background.  Before the table can be swapped with
 * dm_bind_table, dm_suspend must be called to flush any in
 * flight bios and ensure that any further io gets deferred.
 */
int dm_suspend(struct mapped_device *md)
{
	struct dm_table *map;
	DECLARE_WAITQUEUE(wait, current);

	/* Flush I/O to the device. */
	down_read(&md->lock);
	if (test_bit(DMF_BLOCK_IO, &md->flags)) {
		up_read(&md->lock);
		return -EINVAL;
	}

	__lock_fs(md);
	up_read(&md->lock);

	/*
	 * First we set the BLOCK_IO flag so no more ios will be
	 * mapped.
	 */
	down_write(&md->lock);
	if (test_bit(DMF_BLOCK_IO, &md->flags)) {
		/*
		 * If we get here we know another thread is
		 * trying to suspend as well, so we leave the fs
		 * locked for this thread.
		 */
		up_write(&md->lock);
		return -EINVAL;
	}

	set_bit(DMF_BLOCK_IO, &md->flags);
	add_wait_queue(&md->wait, &wait);
	up_write(&md->lock);

	/* unplug */
	map = dm_get_table(md);
	if (map) {
		dm_table_unplug_all(map);
		dm_table_put(map);
	}

	/*
	 * Then we wait for the already mapped ios to
	 * complete.
	 */
	while (1) {
		set_current_state(TASK_INTERRUPTIBLE);

		if (!atomic_read(&md->pending) || signal_pending(current))
			break;

		io_schedule();
	}
	set_current_state(TASK_RUNNING);

	down_write(&md->lock);
	remove_wait_queue(&md->wait, &wait);

	/* were we interrupted ? */
	if (atomic_read(&md->pending)) {
		__unlock_fs(md);
		clear_bit(DMF_BLOCK_IO, &md->flags);
		up_write(&md->lock);
		return -EINTR;
	}

	set_bit(DMF_SUSPENDED, &md->flags);

	map = dm_get_table(md);
	if (map)
		dm_table_suspend_targets(map);
	dm_table_put(map);
	up_write(&md->lock);

	return 0;
}

int dm_resume(struct mapped_device *md)
{
	struct bio *def;
	struct dm_table *map = dm_get_table(md);

	down_write(&md->lock);
	if (!map ||
	    !test_bit(DMF_SUSPENDED, &md->flags) ||
	    !dm_table_get_size(map)) {
		up_write(&md->lock);
		dm_table_put(map);
		return -EINVAL;
	}

	dm_table_resume_targets(map);
	clear_bit(DMF_SUSPENDED, &md->flags);
	clear_bit(DMF_BLOCK_IO, &md->flags);

	def = bio_list_get(&md->deferred);
	__flush_deferred_io(md, def);
	up_write(&md->lock);
	__unlock_fs(md);
	dm_table_unplug_all(map);
	dm_table_put(map);

	return 0;
}

/*-----------------------------------------------------------------
 * Event notification.
 *---------------------------------------------------------------*/
uint32_t dm_get_event_nr(struct mapped_device *md)
{
	return atomic_read(&md->event_nr);
}

int dm_wait_event(struct mapped_device *md, int event_nr)
{
	return wait_event_interruptible(md->eventq,
			(event_nr != atomic_read(&md->event_nr)));
}

/*
 * The gendisk is only valid as long as you have a reference
 * count on 'md'.
 */
struct gendisk *dm_disk(struct mapped_device *md)
{
	return md->disk;
}

int dm_suspended(struct mapped_device *md)
{
	return test_bit(DMF_SUSPENDED, &md->flags);
}

static struct block_device_operations dm_blk_dops = {
	.open = dm_blk_open,
	.release = dm_blk_close,
	.owner = THIS_MODULE
};

/*
 * module hooks
 */
module_init(dm_init);
module_exit(dm_exit);

module_param(major, uint, 0);
MODULE_PARM_DESC(major, "The major number of the device mapper");
MODULE_DESCRIPTION(DM_NAME " driver");
MODULE_AUTHOR("Joe Thornber <[e-mail]>");
MODULE_LICENSE("GPL");