dm-mpath.c

linux 内核源代码

	list_for_each_entry(pg, &m->priority_groups, list) {
		pg->bypassed = 0;
		if (--pgnum)
			continue;

		m->current_pgpath = NULL;
		m->current_pg = NULL;
		m->next_pg = pg;
	}
	spin_unlock_irqrestore(&m->lock, flags);

	queue_work(kmultipathd, &m->trigger_event);
	return 0;
}

/*
 * Set/clear bypassed status of a PG.
 * PGs are numbered upwards from 1 in the order they were declared.
 */
static int bypass_pg_num(struct multipath *m, const char *pgstr, int bypassed)
{
	struct priority_group *pg;
	unsigned pgnum;

	if (!pgstr || (sscanf(pgstr, "%u", &pgnum) != 1) || !pgnum ||
	    (pgnum > m->nr_priority_groups)) {
		DMWARN("invalid PG number supplied to bypass_pg");
		return -EINVAL;
	}

	list_for_each_entry(pg, &m->priority_groups, list) {
		if (!--pgnum)
			break;
	}

	bypass_pg(m, pg, bypassed);
	return 0;
}

/*
 * Should we retry pg_init immediately?
 */
static int pg_init_limit_reached(struct multipath *m, struct pgpath *pgpath)
{
	unsigned long flags;
	int limit_reached = 0;

	spin_lock_irqsave(&m->lock, flags);

	if (m->pg_init_count <= m->pg_init_retries)
		m->pg_init_required = 1;
	else
		limit_reached = 1;

	spin_unlock_irqrestore(&m->lock, flags);

	return limit_reached;
}

/*
 * pg_init must call this when it has completed its initialisation
 */
void dm_pg_init_complete(struct dm_path *path, unsigned err_flags)
{
	struct pgpath *pgpath = path_to_pgpath(path);
	struct priority_group *pg = pgpath->pg;
	struct multipath *m = pg->m;
	unsigned long flags;

	/*
	 * If requested, retry pg_init until maximum number of retries exceeded.
	 * If retry not requested and PG already bypassed, always fail the path.
	 */
	if (err_flags & MP_RETRY) {
		if (pg_init_limit_reached(m, pgpath))
			err_flags |= MP_FAIL_PATH;
	} else if (err_flags && pg->bypassed)
		err_flags |= MP_FAIL_PATH;

	if (err_flags & MP_FAIL_PATH)
		fail_path(pgpath);

	if (err_flags & MP_BYPASS_PG)
		bypass_pg(m, pg, 1);

	spin_lock_irqsave(&m->lock, flags);
	if (err_flags & ~MP_RETRY) {
		m->current_pgpath = NULL;
		m->current_pg = NULL;
	} else if (!m->pg_init_required)
		m->queue_io = 0;

	m->pg_init_in_progress = 0;
	queue_work(kmultipathd, &m->process_queued_ios);
	spin_unlock_irqrestore(&m->lock, flags);
}

/*
 * end_io handling
 */
static int do_end_io(struct multipath *m, struct bio *bio,
		     int error, struct dm_mpath_io *mpio)
{
	struct hw_handler *hwh = &m->hw_handler;
	unsigned err_flags = MP_FAIL_PATH;	/* Default behavior */
	unsigned long flags;

	if (!error)
		return 0;	/* I/O complete */

	if ((error == -EWOULDBLOCK) && bio_rw_ahead(bio))
		return error;

	if (error == -EOPNOTSUPP)
		return error;

	spin_lock_irqsave(&m->lock, flags);
	if (!m->nr_valid_paths) {
		if (__must_push_back(m)) {
			spin_unlock_irqrestore(&m->lock, flags);
			return DM_ENDIO_REQUEUE;
		} else if (!m->queue_if_no_path) {
			spin_unlock_irqrestore(&m->lock, flags);
			return -EIO;
		} else {
			spin_unlock_irqrestore(&m->lock, flags);
			goto requeue;
		}
	}
	spin_unlock_irqrestore(&m->lock, flags);

	if (hwh->type && hwh->type->error)
		err_flags = hwh->type->error(hwh, bio);

	if (mpio->pgpath) {
		if (err_flags & MP_FAIL_PATH)
			fail_path(mpio->pgpath);

		if (err_flags & MP_BYPASS_PG)
			bypass_pg(m, mpio->pgpath->pg, 1);
	}

	if (err_flags & MP_ERROR_IO)
		return -EIO;

      requeue:
	dm_bio_restore(&mpio->details, bio);

	/* queue for the daemon to resubmit or fail */
	spin_lock_irqsave(&m->lock, flags);
	bio_list_add(&m->queued_ios, bio);
	m->queue_size++;
	if (!m->queue_io)
		queue_work(kmultipathd, &m->process_queued_ios);
	spin_unlock_irqrestore(&m->lock, flags);

	return DM_ENDIO_INCOMPLETE;	/* io not complete */
}

static int multipath_end_io(struct dm_target *ti, struct bio *bio,
			    int error, union map_info *map_context)
{
	struct multipath *m = ti->private;
	struct dm_mpath_io *mpio = map_context->ptr;
	struct pgpath *pgpath = mpio->pgpath;
	struct path_selector *ps;
	int r;

	r  = do_end_io(m, bio, error, mpio);
	if (pgpath) {
		ps = &pgpath->pg->ps;
		if (ps->type->end_io)
			ps->type->end_io(ps, &pgpath->path);
	}
	if (r != DM_ENDIO_INCOMPLETE)
		mempool_free(mpio, m->mpio_pool);

	return r;
}

/*
 * Suspend can't complete until all the I/O is processed so if
 * the last path fails we must error any remaining I/O.
 * Note that if the freeze_bdev fails while suspending, the
 * queue_if_no_path state is lost - userspace should reset it.
 */
static void multipath_presuspend(struct dm_target *ti)
{
	struct multipath *m = (struct multipath *) ti->private;

	queue_if_no_path(m, 0, 1);
}

/*
 * Restore the queue_if_no_path setting.
 */
static void multipath_resume(struct dm_target *ti)
{
	struct multipath *m = (struct multipath *) ti->private;
	unsigned long flags;

	spin_lock_irqsave(&m->lock, flags);
	m->queue_if_no_path = m->saved_queue_if_no_path;
	spin_unlock_irqrestore(&m->lock, flags);
}

/*
 * Info output has the following format:
 * num_multipath_feature_args [multipath_feature_args]*
 * num_handler_status_args [handler_status_args]*
 * num_groups init_group_number
 *            [A|D|E num_ps_status_args [ps_status_args]*
 *             num_paths num_selector_args
 *             [path_dev A|F fail_count [selector_args]* ]+ ]+
 *
 * Table output has the following format (identical to the constructor string):
 * num_feature_args [features_args]*
 * num_handler_args hw_handler [hw_handler_args]*
 * num_groups init_group_number
 *     [priority selector-name num_ps_args [ps_args]*
 *      num_paths num_selector_args [path_dev [selector_args]* ]+ ]+
 */
static int multipath_status(struct dm_target *ti, status_type_t type,
			    char *result, unsigned int maxlen)
{
	int sz = 0;
	unsigned long flags;
	struct multipath *m = (struct multipath *) ti->private;
	struct hw_handler *hwh = &m->hw_handler;
	struct priority_group *pg;
	struct pgpath *p;
	unsigned pg_num;
	char state;

	spin_lock_irqsave(&m->lock, flags);

	/* Features */
	if (type == STATUSTYPE_INFO)
		DMEMIT("2 %u %u ", m->queue_size, m->pg_init_count);
	else {
		DMEMIT("%u ", m->queue_if_no_path +
			      (m->pg_init_retries > 0) * 2);
		if (m->queue_if_no_path)
			DMEMIT("queue_if_no_path ");
		if (m->pg_init_retries)
			DMEMIT("pg_init_retries %u ", m->pg_init_retries);
	}

	if (hwh->type && hwh->type->status)
		sz += hwh->type->status(hwh, type, result + sz, maxlen - sz);
	else if (!hwh->type || type == STATUSTYPE_INFO)
		DMEMIT("0 ");
	else
		DMEMIT("1 %s ", hwh->type->name);

	DMEMIT("%u ", m->nr_priority_groups);

	if (m->next_pg)
		pg_num = m->next_pg->pg_num;
	else if (m->current_pg)
		pg_num = m->current_pg->pg_num;
	else
			pg_num = 1;

	DMEMIT("%u ", pg_num);

	switch (type) {
	case STATUSTYPE_INFO:
		list_for_each_entry(pg, &m->priority_groups, list) {
			if (pg->bypassed)
				state = 'D';	/* Disabled */
			else if (pg == m->current_pg)
				state = 'A';	/* Currently Active */
			else
				state = 'E';	/* Enabled */

			DMEMIT("%c ", state);

			if (pg->ps.type->status)
				sz += pg->ps.type->status(&pg->ps, NULL, type,
							  result + sz,
							  maxlen - sz);
			else
				DMEMIT("0 ");

			DMEMIT("%u %u ", pg->nr_pgpaths,
			       pg->ps.type->info_args);

			list_for_each_entry(p, &pg->pgpaths, list) {
				DMEMIT("%s %s %u ", p->path.dev->name,
				       p->path.is_active ? "A" : "F",
				       p->fail_count);
				if (pg->ps.type->status)
					sz += pg->ps.type->status(&pg->ps,
					      &p->path, type, result + sz,
					      maxlen - sz);
			}
		}
		break;

	case STATUSTYPE_TABLE:
		list_for_each_entry(pg, &m->priority_groups, list) {
			DMEMIT("%s ", pg->ps.type->name);

			if (pg->ps.type->status)
				sz += pg->ps.type->status(&pg->ps, NULL, type,
							  result + sz,
							  maxlen - sz);
			else
				DMEMIT("0 ");

			DMEMIT("%u %u ", pg->nr_pgpaths,
			       pg->ps.type->table_args);

			list_for_each_entry(p, &pg->pgpaths, list) {
				DMEMIT("%s ", p->path.dev->name);
				if (pg->ps.type->status)
					sz += pg->ps.type->status(&pg->ps,
					      &p->path, type, result + sz,
					      maxlen - sz);
			}
		}
		break;
	}

	spin_unlock_irqrestore(&m->lock, flags);

	return 0;
}

static int multipath_message(struct dm_target *ti, unsigned argc, char **argv)
{
	int r;
	struct dm_dev *dev;
	struct multipath *m = (struct multipath *) ti->private;
	action_fn action;

	if (argc == 1) {
		if (!strnicmp(argv[0], MESG_STR("queue_if_no_path")))
			return queue_if_no_path(m, 1, 0);
		else if (!strnicmp(argv[0], MESG_STR("fail_if_no_path")))
			return queue_if_no_path(m, 0, 0);
	}

	if (argc != 2)
		goto error;

	if (!strnicmp(argv[0], MESG_STR("disable_group")))
		return bypass_pg_num(m, argv[1], 1);
	else if (!strnicmp(argv[0], MESG_STR("enable_group")))
		return bypass_pg_num(m, argv[1], 0);
	else if (!strnicmp(argv[0], MESG_STR("switch_group")))
		return switch_pg_num(m, argv[1]);
	else if (!strnicmp(argv[0], MESG_STR("reinstate_path")))
		action = reinstate_path;
	else if (!strnicmp(argv[0], MESG_STR("fail_path")))
		action = fail_path;
	else
		goto error;

	r = dm_get_device(ti, argv[1], ti->begin, ti->len,
			  dm_table_get_mode(ti->table), &dev);
	if (r) {
		DMWARN("message: error getting device %s",
		       argv[1]);
		return -EINVAL;
	}

	r = action_dev(m, dev, action);

	dm_put_device(ti, dev);

	return r;

error:
	DMWARN("Unrecognised multipath message received.");
	return -EINVAL;
}

static int multipath_ioctl(struct dm_target *ti, struct inode *inode,
			   struct file *filp, unsigned int cmd,
			   unsigned long arg)
{
	struct multipath *m = (struct multipath *) ti->private;
	struct block_device *bdev = NULL;
	unsigned long flags;
	struct file fake_file = {};
	struct dentry fake_dentry = {};
	int r = 0;

	fake_file.f_path.dentry = &fake_dentry;

	spin_lock_irqsave(&m->lock, flags);

	if (!m->current_pgpath)
		__choose_pgpath(m);

	if (m->current_pgpath) {
		bdev = m->current_pgpath->path.dev->bdev;
		fake_dentry.d_inode = bdev->bd_inode;
		fake_file.f_mode = m->current_pgpath->path.dev->mode;
	}

	if (m->queue_io)
		r = -EAGAIN;
	else if (!bdev)
		r = -EIO;

	spin_unlock_irqrestore(&m->lock, flags);

	return r ? : blkdev_driver_ioctl(bdev->bd_inode, &fake_file,
					 bdev->bd_disk, cmd, arg);
}

/*-----------------------------------------------------------------
 * Module setup
 *---------------------------------------------------------------*/
static struct target_type multipath_target = {
	.name = "multipath",
	.version = {1, 0, 5},
	.module = THIS_MODULE,
	.ctr = multipath_ctr,
	.dtr = multipath_dtr,
	.map = multipath_map,
	.end_io = multipath_end_io,
	.presuspend = multipath_presuspend,
	.resume = multipath_resume,
	.status = multipath_status,
	.message = multipath_message,
	.ioctl  = multipath_ioctl,
};

static int __init dm_multipath_init(void)
{
	int r;

	/* allocate a slab for the dm_ios */
	_mpio_cache = KMEM_CACHE(dm_mpath_io, 0);
	if (!_mpio_cache)
		return -ENOMEM;

	r = dm_register_target(&multipath_target);
	if (r < 0) {
		DMERR("register failed %d", r);
		kmem_cache_destroy(_mpio_cache);
		return -EINVAL;
	}

	kmultipathd = create_workqueue("kmpathd");
	if (!kmultipathd) {
		DMERR("failed to create workqueue kmpathd");
		dm_unregister_target(&multipath_target);
		kmem_cache_destroy(_mpio_cache);
		return -ENOMEM;
	}

	DMINFO("version %u.%u.%u loaded",
	       multipath_target.version[0], multipath_target.version[1],
	       multipath_target.version[2]);

	return r;
}

static void __exit dm_multipath_exit(void)
{
	int r;

	destroy_workqueue(kmultipathd);

	r = dm_unregister_target(&multipath_target);
	if (r < 0)
		DMERR("target unregister failed %d", r);
	kmem_cache_destroy(_mpio_cache);
}

EXPORT_SYMBOL_GPL(dm_pg_init_complete);

module_init(dm_multipath_init);
module_exit(dm_multipath_exit);

MODULE_DESCRIPTION(DM_NAME " multipath target");
MODULE_AUTHOR("Sistina Software <[e-mail]>");
MODULE_LICENSE("GPL");