drivers/md/md.c

// SPDX-License-Identifier: GPL-2.0-or-later
/*
   md.c : Multiple Devices driver for Linux
     Copyright (C) 1998, 1999, 2000 Ingo Molnar

     completely rewritten, based on the MD driver code from Marc Zyngier

   Changes:

   - RAID-1/RAID-5 extensions by Miguel de Icaza, Gadi Oxman, Ingo Molnar
   - RAID-6 extensions by H. Peter Anvin <hpa@zytor.com>
   - boot support for linear and striped mode by Harald Hoyer <HarryH@Royal.Net>
   - kerneld support by Boris Tobotras <boris@xtalk.msk.su>
   - kmod support by: Cyrus Durgin
   - RAID0 bugfixes: Mark Anthony Lisher <markal@iname.com>
   - Devfs support by Richard Gooch <rgooch@atnf.csiro.au>

   - lots of fixes and improvements to the RAID1/RAID5 and generic
     RAID code (such as request based resynchronization):

     Neil Brown <neilb@cse.unsw.edu.au>.

   - persistent bitmap code
     Copyright (C) 2003-2004, Paul Clements, SteelEye Technology, Inc.


   Errors, Warnings, etc.
   Please use:
     pr_crit() for error conditions that risk data loss
     pr_err() for error conditions that are unexpected, like an IO error
         or internal inconsistency
     pr_warn() for error conditions that could have been predicated, like
         adding a device to an array when it has incompatible metadata
     pr_info() for every interesting, very rare events, like an array starting
         or stopping, or resync starting or stopping
     pr_debug() for everything else.

*/

#include <linux/sched/mm.h>
#include <linux/sched/signal.h>
#include <linux/kthread.h>
#include <linux/blkdev.h>
#include <linux/blk-integrity.h>
#include <linux/badblocks.h>
#include <linux/sysctl.h>
#include <linux/seq_file.h>
#include <linux/fs.h>
#include <linux/poll.h>
#include <linux/ctype.h>
#include <linux/string.h>
#include <linux/hdreg.h>
#include <linux/proc_fs.h>
#include <linux/random.h>
#include <linux/major.h>
#include <linux/module.h>
#include <linux/reboot.h>
#include <linux/file.h>
#include <linux/compat.h>
#include <linux/delay.h>
#include <linux/raid/md_p.h>
#include <linux/raid/md_u.h>
#include <linux/raid/detect.h>
#include <linux/slab.h>
#include <linux/percpu-refcount.h>
#include <linux/part_stat.h>

#include "md.h"
#include "md-bitmap.h"
#include "md-cluster.h"

static const char *action_name[NR_SYNC_ACTIONS] = {
	[ACTION_RESYNC]		= "resync",
	[ACTION_RECOVER]	= "recover",
	[ACTION_CHECK]		= "check",
	[ACTION_REPAIR]		= "repair",
	[ACTION_RESHAPE]	= "reshape",
	[ACTION_FROZEN]		= "frozen",
	[ACTION_IDLE]		= "idle",
};

/* pers_list is a list of registered personalities protected by pers_lock. */
static LIST_HEAD(pers_list);
static DEFINE_SPINLOCK(pers_lock);

static const struct kobj_type md_ktype;

const struct md_cluster_operations *md_cluster_ops;
EXPORT_SYMBOL(md_cluster_ops);
static struct module *md_cluster_mod;

static DECLARE_WAIT_QUEUE_HEAD(resync_wait);
static struct workqueue_struct *md_wq;

/*
 * This workqueue is used for sync_work to register new sync_thread, and for
 * del_work to remove rdev, and for event_work that is only set by dm-raid.
 *
 * Noted that sync_work will grab reconfig_mutex, hence never flush this
 * workqueue whith reconfig_mutex grabbed.
 */
static struct workqueue_struct *md_misc_wq;
struct workqueue_struct *md_bitmap_wq;

static int remove_and_add_spares(struct mddev *mddev,
				 struct md_rdev *this);
static void mddev_detach(struct mddev *mddev);
static void export_rdev(struct md_rdev *rdev, struct mddev *mddev);
static void md_wakeup_thread_directly(struct md_thread __rcu *thread);

/*
 * Default number of read corrections we'll attempt on an rdev
 * before ejecting it from the array. We divide the read error
 * count by 2 for every hour elapsed between read errors.
 */
#define MD_DEFAULT_MAX_CORRECTED_READ_ERRORS 20
/* Default safemode delay: 200 msec */
#define DEFAULT_SAFEMODE_DELAY ((200 * HZ)/1000 +1)
/*
 * Current RAID-1,4,5 parallel reconstruction 'guaranteed speed limit'
 * is 1000 KB/sec, so the extra system load does not show up that much.
 * Increase it if you want to have more _guaranteed_ speed. Note that
 * the RAID driver will use the maximum available bandwidth if the IO
 * subsystem is idle. There is also an 'absolute maximum' reconstruction
 * speed limit - in case reconstruction slows down your system despite
 * idle IO detection.
 *
 * you can change it via /proc/sys/dev/raid/speed_limit_min and _max.
 * or /sys/block/mdX/md/sync_speed_{min,max}
 */

static int sysctl_speed_limit_min = 1000;
static int sysctl_speed_limit_max = 200000;
static inline int speed_min(struct mddev *mddev)
{
	return mddev->sync_speed_min ?
		mddev->sync_speed_min : sysctl_speed_limit_min;
}

static inline int speed_max(struct mddev *mddev)
{
	return mddev->sync_speed_max ?
		mddev->sync_speed_max : sysctl_speed_limit_max;
}

static void rdev_uninit_serial(struct md_rdev *rdev)
{
	if (!test_and_clear_bit(CollisionCheck, &rdev->flags))
		return;

	kvfree(rdev->serial);
	rdev->serial = NULL;
}

static void rdevs_uninit_serial(struct mddev *mddev)
{
	struct md_rdev *rdev;

	rdev_for_each(rdev, mddev)
		rdev_uninit_serial(rdev);
}

static int rdev_init_serial(struct md_rdev *rdev)
{
	/* serial_nums equals with BARRIER_BUCKETS_NR */
	int i, serial_nums = 1 << ((PAGE_SHIFT - ilog2(sizeof(atomic_t))));
	struct serial_in_rdev *serial = NULL;

	if (test_bit(CollisionCheck, &rdev->flags))
		return 0;

	serial = kvmalloc(sizeof(struct serial_in_rdev) * serial_nums,
			  GFP_KERNEL);
	if (!serial)
		return -ENOMEM;

	for (i = 0; i < serial_nums; i++) {
		struct serial_in_rdev *serial_tmp = &serial[i];

		spin_lock_init(&serial_tmp->serial_lock);
		serial_tmp->serial_rb = RB_ROOT_CACHED;
		init_waitqueue_head(&serial_tmp->serial_io_wait);
	}

	rdev->serial = serial;
	set_bit(CollisionCheck, &rdev->flags);

	return 0;
}

static int rdevs_init_serial(struct mddev *mddev)
{
	struct md_rdev *rdev;
	int ret = 0;

	rdev_for_each(rdev, mddev) {
		ret = rdev_init_serial(rdev);
		if (ret)
			break;
	}

	/* Free all resources if pool is not existed */
	if (ret && !mddev->serial_info_pool)
		rdevs_uninit_serial(mddev);

	return ret;
}

/*
 * rdev needs to enable serial stuffs if it meets the conditions:
 * 1. it is multi-queue device flaged with writemostly.
 * 2. the write-behind mode is enabled.
 */
static int rdev_need_serial(struct md_rdev *rdev)
{
	return (rdev && rdev->mddev->bitmap_info.max_write_behind > 0 &&
		rdev->bdev->bd_disk->queue->nr_hw_queues != 1 &&
		test_bit(WriteMostly, &rdev->flags));
}

/*
 * Init resource for rdev(s), then create serial_info_pool if:
 * 1. rdev is the first device which return true from rdev_enable_serial.
 * 2. rdev is NULL, means we want to enable serialization for all rdevs.
 */
void mddev_create_serial_pool(struct mddev *mddev, struct md_rdev *rdev)
{
	int ret = 0;

	if (rdev && !rdev_need_serial(rdev) &&
	    !test_bit(CollisionCheck, &rdev->flags))
		return;

	if (!rdev)
		ret = rdevs_init_serial(mddev);
	else
		ret = rdev_init_serial(rdev);
	if (ret)
		return;

	if (mddev->serial_info_pool == NULL) {
		/*
		 * already in memalloc noio context by
		 * mddev_suspend()
		 */
		mddev->serial_info_pool =
			mempool_create_kmalloc_pool(NR_SERIAL_INFOS,
						sizeof(struct serial_info));
		if (!mddev->serial_info_pool) {
			rdevs_uninit_serial(mddev);
			pr_err("can't alloc memory pool for serialization\n");
		}
	}
}

/*
 * Free resource from rdev(s), and destroy serial_info_pool under conditions:
 * 1. rdev is the last device flaged with CollisionCheck.
 * 2. when bitmap is destroyed while policy is not enabled.
 * 3. for disable policy, the pool is destroyed only when no rdev needs it.
 */
void mddev_destroy_serial_pool(struct mddev *mddev, struct md_rdev *rdev)
{
	if (rdev && !test_bit(CollisionCheck, &rdev->flags))
		return;

	if (mddev->serial_info_pool) {
		struct md_rdev *temp;
		int num = 0; /* used to track if other rdevs need the pool */

		rdev_for_each(temp, mddev) {
			if (!rdev) {
				if (!mddev->serialize_policy ||
				    !rdev_need_serial(temp))
					rdev_uninit_serial(temp);
				else
					num++;
			} else if (temp != rdev &&
				   test_bit(CollisionCheck, &temp->flags))
				num++;
		}

		if (rdev)
			rdev_uninit_serial(rdev);

		if (num)
			pr_info("The mempool could be used by other devices\n");
		else {
			mempool_destroy(mddev->serial_info_pool);
			mddev->serial_info_pool = NULL;
		}
	}
}

static struct ctl_table_header *raid_table_header;

static struct ctl_table raid_table[] = {
	{
		.procname	= "speed_limit_min",
		.data		= &sysctl_speed_limit_min,
		.maxlen		= sizeof(int),
		.mode		= S_IRUGO|S_IWUSR,
		.proc_handler	= proc_dointvec,
	},
	{
		.procname	= "speed_limit_max",
		.data		= &sysctl_speed_limit_max,
		.maxlen		= sizeof(int),
		.mode		= S_IRUGO|S_IWUSR,
		.proc_handler	= proc_dointvec,
	},
};

static int start_readonly;

/*
 * The original mechanism for creating an md device is to create
 * a device node in /dev and to open it.  This causes races with device-close.
 * The preferred method is to write to the "new_array" module parameter.
 * This can avoid races.
 * Setting create_on_open to false disables the original mechanism
 * so all the races disappear.
 */
static bool create_on_open = true;

/*
 * We have a system wide 'event count' that is incremented
 * on any 'interesting' event, and readers of /proc/mdstat
 * can use 'poll' or 'select' to find out when the event
 * count increases.
 *
 * Events are:
 *  start array, stop array, error, add device, remove device,
 *  start build, activate spare
 */
static DECLARE_WAIT_QUEUE_HEAD(md_event_waiters);
static atomic_t md_event_count;
void md_new_event(void)
{
	atomic_inc(&md_event_count);
	wake_up(&md_event_waiters);
}
EXPORT_SYMBOL_GPL(md_new_event);

/*
 * Enables to iterate over all existing md arrays
 * all_mddevs_lock protects this list.
 */
static LIST_HEAD(all_mddevs);
static DEFINE_SPINLOCK(all_mddevs_lock);

static bool is_md_suspended(struct mddev *mddev)
{
	return percpu_ref_is_dying(&mddev->active_io);
}
/* Rather than calling directly into the personality make_request function,
 * IO requests come here first so that we can check if the device is
 * being suspended pending a reconfiguration.
 * We hold a refcount over the call to ->make_request.  By the time that
 * call has finished, the bio has been linked into some internal structure
 * and so is visible to ->quiesce(), so we don't need the refcount any more.
 */
static bool is_suspended(struct mddev *mddev, struct bio *bio)
{
	if (is_md_suspended(mddev))
		return true;
	if (bio_data_dir(bio) != WRITE)
		return false;
	if (READ_ONCE(mddev->suspend_lo) >= READ_ONCE(mddev->suspend_hi))
		return false;
	if (bio->bi_iter.bi_sector >= READ_ONCE(mddev->suspend_hi))
		return false;
	if (bio_end_sector(bio) < READ_ONCE(mddev->suspend_lo))
		return false;
	return true;
}

bool md_handle_request(struct mddev *mddev, struct bio *bio)
{
check_suspended:
	if (is_suspended(mddev, bio)) {
		DEFINE_WAIT(__wait);
		/* Bail out if REQ_NOWAIT is set for the bio */
		if (bio->bi_opf & REQ_NOWAIT) {
			bio_wouldblock_error(bio);
			return true;
		}
		for (;;) {
			prepare_to_wait(&mddev->sb_wait, &__wait,
					TASK_UNINTERRUPTIBLE);
			if (!is_suspended(mddev, bio))
				break;
			schedule();
		}
		finish_wait(&mddev->sb_wait, &__wait);
	}
	if (!percpu_ref_tryget_live(&mddev->active_io))
		goto check_suspended;

	if (!mddev->pers->make_request(mddev, bio)) {
		percpu_ref_put(&mddev->active_io);
		if (!mddev->gendisk && mddev->pers->prepare_suspend)
			return false;
		goto check_suspended;
	}

	percpu_ref_put(&mddev->active_io);
	return true;
}
EXPORT_SYMBOL(md_handle_request);

static void md_submit_bio(struct bio *bio)
{
	const int rw = bio_data_dir(bio);
	struct mddev *mddev = bio->bi_bdev->bd_disk->private_data;

	if (mddev == NULL || mddev->pers == NULL) {
		bio_io_error(bio);
		return;
	}

	if (unlikely(test_bit(MD_BROKEN, &mddev->flags)) && (rw == WRITE)) {
		bio_io_error(bio);
		return;
	}

	bio = bio_split_to_limits(bio);
	if (!bio)
		return;

	if (mddev->ro == MD_RDONLY && unlikely(rw == WRITE)) {
		if (bio_sectors(bio) != 0)
			bio->bi_status = BLK_STS_IOERR;
		bio_endio(bio);
		return;
	}

	/* bio could be mergeable after passing to underlayer */
	bio->bi_opf &= ~REQ_NOMERGE;

	md_handle_request(mddev, bio);
}

/*
 * Make sure no new requests are submitted to the device, and any requests that
 * have been submitted are completely handled.
 */
int mddev_suspend(struct mddev *mddev, bool interruptible)
{
	int err = 0;

	/*
	 * hold reconfig_mutex to wait for normal io will deadlock, because
	 * other context can't update super_block, and normal io can rely on
	 * updating super_block.
	 */
	lockdep_assert_not_held(&mddev->reconfig_mutex);

	if (interruptible)
		err = mutex_lock_interruptible(&mddev->suspend_mutex);
	else
		mutex_lock(&mddev->suspend_mutex);
	if (err)
		return err;

	if (mddev->suspended) {
		WRITE_ONCE(mddev->suspended, mddev->suspended + 1);
		mutex_unlock(&mddev->suspend_mutex);
		return 0;
	}

	percpu_ref_kill(&mddev->active_io);
	if (interruptible)
		err = wait_event_interruptible(mddev->sb_wait,
				percpu_ref_is_zero(&mddev->active_io));
	else
		wait_event(mddev->sb_wait,
				percpu_ref_is_zero(&mddev->active_io));
	if (err) {
		percpu_ref_resurrect(&mddev->active_io);
		mutex_unlock(&mddev->suspend_mutex);
		return err;
	}

	/*
	 * For raid456, io might be waiting for reshape to make progress,
	 * allow new reshape to start while waiting for io to be done to
	 * prevent deadlock.
	 */
	WRITE_ONCE(mddev->suspended, mddev->suspended + 1);

	/* restrict memory reclaim I/O during raid array is suspend */
	mddev->noio_flag = memalloc_noio_save();

	mutex_unlock(&mddev->suspend_mutex);
	return 0;
}
EXPORT_SYMBOL_GPL(mddev_suspend);

static void __mddev_resume(struct mddev *mddev, bool recovery_needed)
{
	lockdep_assert_not_held(&mddev->reconfig_mutex);

	mutex_lock(&mddev->suspend_mutex);
	WRITE_ONCE(mddev->suspended, mddev->suspended - 1);
	if (mddev->suspended) {
		mutex_unlock(&mddev->suspend_mutex);
		return;
	}

	/* entred the memalloc scope from mddev_suspend() */
	memalloc_noio_restore(mddev->noio_flag);

	percpu_ref_resurrect(&mddev->active_io);
	wake_up(&mddev->sb_wait);

	if (recovery_needed)
		set_bit(MD_RECOVERY_NEEDED, &mddev->recovery);
	md_wakeup_thread(mddev->thread);
	md_wakeup_thread(mddev->sync_thread); /* possibly kick off a reshape */

	mutex_unlock(&mddev->suspend_mutex);
}

void mddev_resume(struct mddev *mddev)
{
	return __mddev_resume(mddev, true);
}
EXPORT_SYMBOL_GPL(mddev_resume);

/* sync bdev before setting device to readonly or stopping raid*/
static int mddev_set_closing_and_sync_blockdev(struct mddev *mddev, int opener_num)
{
	mutex_lock(&mddev->open_mutex);
	if (mddev->pers && atomic_read(&mddev->openers) > opener_num) {
		mutex_unlock(&mddev->open_mutex);
		return -EBUSY;
	}
	if (test_and_set_bit(MD_CLOSING, &mddev->flags)) {
		mutex_unlock(&mddev->open_mutex);
		return -EBUSY;
	}
	mutex_unlock(&mddev->open_mutex);

	sync_blockdev(mddev->gendisk->part0);
	return 0;
}

/*
 * Generic flush handling for md
 */

static void md_end_flush(struct bio *bio)
{
	struct md_rdev *rdev = bio->bi_private;
	struct mddev *mddev = rdev->mddev;

	bio_put(bio);

	rdev_dec_pending(rdev, mddev);

	if (atomic_dec_and_test(&mddev->flush_pending))
		/* The pre-request flush has finished */
		queue_work(md_wq, &mddev->flush_work);
}

static void md_submit_flush_data(struct work_struct *ws);

static void submit_flushes(struct work_struct *ws)
{
	struct mddev *mddev = container_of(ws, struct mddev, flush_work);
	struct md_rdev *rdev;

	mddev->start_flush = ktime_get_boottime();
	INIT_WORK(&mddev->flush_work, md_submit_flush_data);
	atomic_set(&mddev->flush_pending, 1);
	rcu_read_lock();
	rdev_for_each_rcu(rdev, mddev)
		if (rdev->raid_disk >= 0 &&
		    !test_bit(Faulty, &rdev->flags)) {
			struct bio *bi;

			atomic_inc(&rdev->nr_pending);
			rcu_read_unlock();
			bi = bio_alloc_bioset(rdev->bdev, 0,
					      REQ_OP_WRITE | REQ_PREFLUSH,
					      GFP_NOIO, &mddev->bio_set);
			bi->bi_end_io = md_end_flush;
			bi->bi_private = rdev;
			atomic_inc(&mddev->flush_pending);
			submit_bio(bi);
			rcu_read_lock();
		}
	rcu_read_unlock();
	if (atomic_dec_and_test(&mddev->flush_pending))
		queue_work(md_wq, &mddev->flush_work);
}

static void md_submit_flush_data(struct work_struct *ws)
{
	struct mddev *mddev = container_of(ws, struct mddev, flush_work);
	struct bio *bio = mddev->flush_bio;

	/*
	 * must reset flush_bio before calling into md_handle_request to avoid a
	 * deadlock, because other bios passed md_handle_request suspend check
	 * could wait for this and below md_handle_request could wait for those
	 * bios because of suspend check
	 */
	spin_lock_irq(&mddev->lock);
	mddev->prev_flush_start = mddev->start_flush;
	mddev->flush_bio = NULL;
	spin_unlock_irq(&mddev->lock);
	wake_up(&mddev->sb_wait);

	if (bio->bi_iter.bi_size == 0) {
		/* an empty barrier - all done */
		bio_endio(bio);
	} else {
		bio->bi_opf &= ~REQ_PREFLUSH;

		/*
		 * make_requst() will never return error here, it only
		 * returns error in raid5_make_request() by dm-raid.
		 * Since dm always splits data and flush operation into
		 * two separate io, io size of flush submitted by dm
		 * always is 0, make_request() will not be called here.
		 */
		if (WARN_ON_ONCE(!mddev->pers->make_request(mddev, bio)))
			bio_io_error(bio);
	}

	/* The pair is percpu_ref_get() from md_flush_request() */
	percpu_ref_put(&mddev->active_io);
}

/*
 * Manages consolidation of flushes and submitting any flushes needed for
 * a bio with REQ_PREFLUSH.  Returns true if the bio is finished or is
 * being finished in another context.  Returns false if the flushing is
 * complete but still needs the I/O portion of the bio to be processed.
 */
bool md_flush_request(struct mddev *mddev, struct bio *bio)
{
	ktime_t req_start = ktime_get_boottime();
	spin_lock_irq(&mddev->lock);
	/* flush requests wait until ongoing flush completes,
	 * hence coalescing all the pending requests.
	 */
	wait_event_lock_irq(mddev->sb_wait,
			    !mddev->flush_bio ||
			    ktime_before(req_start, mddev->prev_flush_start),
			    mddev->lock);
	/* new request after previous flush is completed */
	if (ktime_after(req_start, mddev->prev_flush_start)) {
		WARN_ON(mddev->flush_bio);
		/*
		 * Grab a reference to make sure mddev_suspend() will wait for
		 * this flush to be done.
		 *
		 * md_flush_reqeust() is called under md_handle_request() and
		 * 'active_io' is already grabbed, hence percpu_ref_is_zero()
		 * won't pass, percpu_ref_tryget_live() can't be used because
		 * percpu_ref_kill() can be called by mddev_suspend()
		 * concurrently.
		 */
		WARN_ON(percpu_ref_is_zero(&mddev->active_io));
		percpu_ref_get(&mddev->active_io);
		mddev->flush_bio = bio;
		spin_unlock_irq(&mddev->lock);
		INIT_WORK(&mddev->flush_work, submit_flushes);
		queue_work(md_wq, &mddev->flush_work);
		return true;
	}

	/* flush was performed for some other bio while we waited. */
	spin_unlock_irq(&mddev->lock);
	if (bio->bi_iter.bi_size == 0) {
		/* pure flush without data - all done */
		bio_endio(bio);
		return true;
	}

	bio->bi_opf &= ~REQ_PREFLUSH;
	return false;
}
EXPORT_SYMBOL(md_flush_request);

static inline struct mddev *mddev_get(struct mddev *mddev)
{
	lockdep_assert_held(&all_mddevs_lock);

	if (test_bit(MD_DELETED, &mddev->flags))
		return NULL;
	atomic_inc(&mddev->active);
	return mddev;
}

static void mddev_delayed_delete(struct work_struct *ws);

static void __mddev_put(struct mddev *mddev)
{
	if (mddev->raid_disks || !list_empty(&mddev->disks) ||
	    mddev->ctime || mddev->hold_active)
		return;

	/* Array is not configured at all, and not held active, so destroy it */
	set_bit(MD_DELETED, &mddev->flags);

	/*
	 * Call queue_work inside the spinlock so that flush_workqueue() after
	 * mddev_find will succeed in waiting for the work to be done.
	 */
	queue_work(md_misc_wq, &mddev->del_work);
}

void mddev_put(struct mddev *mddev)
{
	if (!atomic_dec_and_lock(&mddev->active, &all_mddevs_lock))
		return;

	__mddev_put(mddev);
	spin_unlock(&all_mddevs_lock);
}

static void md_safemode_timeout(struct timer_list *t);
static void md_start_sync(struct work_struct *ws);

static void active_io_release(struct percpu_ref *ref)
{
	struct mddev *mddev = container_of(ref, struct mddev, active_io);

	wake_up(&mddev->sb_wait);
}

static void no_op(struct percpu_ref *r) {}

int mddev_init(struct mddev *mddev)
{

	if (percpu_ref_init(&mddev->active_io, active_io_release,
			    PERCPU_REF_ALLOW_REINIT, GFP_KERNEL))
		return -ENOMEM;

	if (percpu_ref_init(&mddev->writes_pending, no_op,
			    PERCPU_REF_ALLOW_REINIT, GFP_KERNEL)) {
		percpu_ref_exit(&mddev->active_io);
		return -ENOMEM;
	}

	/* We want to start with the refcount at zero */
	percpu_ref_put(&mddev->writes_pending);

	mutex_init(&mddev->open_mutex);
	mutex_init(&mddev->reconfig_mutex);
	mutex_init(&mddev->suspend_mutex);
	mutex_init(&mddev->bitmap_info.mutex);
	INIT_LIST_HEAD(&mddev->disks);
	INIT_LIST_HEAD(&mddev->all_mddevs);
	INIT_LIST_HEAD(&mddev->deleting);
	timer_setup(&mddev->safemode_timer, md_safemode_timeout, 0);
	atomic_set(&mddev->active, 1);
	atomic_set(&mddev->openers, 0);
	atomic_set(&mddev->sync_seq, 0);
	spin_lock_init(&mddev->lock);
	atomic_set(&mddev->flush_pending, 0);
	init_waitqueue_head(&mddev->sb_wait);
	init_waitqueue_head(&mddev->recovery_wait);
	mddev->reshape_position = MaxSector;
	mddev->reshape_backwards = 0;
	mddev->last_sync_action = ACTION_IDLE;
	mddev->resync_min = 0;
	mddev->resync_max = MaxSector;
	mddev->level = LEVEL_NONE;

	INIT_WORK(&mddev->sync_work, md_start_sync);
	INIT_WORK(&mddev->del_work, mddev_delayed_delete);

	return 0;
}
EXPORT_SYMBOL_GPL(mddev_init);

void mddev_destroy(struct mddev *mddev)
{
	percpu_ref_exit(&mddev->active_io);
	percpu_ref_exit(&mddev->writes_pending);
}
EXPORT_SYMBOL_GPL(mddev_destroy);

static struct mddev *mddev_find_locked(dev_t unit)
{
	struct mddev *mddev;

	list_for_each_entry(mddev, &all_mddevs, all_mddevs)
		if (mddev->unit == unit)
			return mddev;

	return NULL;
}

/* find an unused unit number */
static dev_t mddev_alloc_unit(void)
{
	static int next_minor = 512;
	int start = next_minor;
	bool is_free = 0;
	dev_t dev = 0;

	while (!is_free) {
		dev = MKDEV(MD_MAJOR, next_minor);
		next_minor++;
		if (next_minor > MINORMASK)
			next_minor = 0;
		if (next_minor == start)
			return 0;		/* Oh dear, all in use. */
		is_free = !mddev_find_locked(dev);
	}

	return dev;
}

static struct mddev *mddev_alloc(dev_t unit)
{
	struct mddev *new;
	int error;

	if (unit && MAJOR(unit) != MD_MAJOR)
		unit &= ~((1 << MdpMinorShift) - 1);

	new = kzalloc(sizeof(*new), GFP_KERNEL);
	if (!new)
		return ERR_PTR(-ENOMEM);

	error = mddev_init(new);
	if (error)
		goto out_free_new;

	spin_lock(&all_mddevs_lock);
	if (unit) {
		error = -EEXIST;
		if (mddev_find_locked(unit))
			goto out_destroy_new;
		new->unit = unit;
		if (MAJOR(unit) == MD_MAJOR)
			new->md_minor = MINOR(unit);
		else
			new->md_minor = MINOR(unit) >> MdpMinorShift;
		new->hold_active = UNTIL_IOCTL;
	} else {
		error = -ENODEV;
		new->unit = mddev_alloc_unit();
		if (!new->unit)
			goto out_destroy_new;
		new->md_minor = MINOR(new->unit);
		new->hold_active = UNTIL_STOP;
	}

	list_add(&new->all_mddevs, &all_mddevs);
	spin_unlock(&all_mddevs_lock);
	return new;

out_destroy_new:
	spin_unlock(&all_mddevs_lock);
	mddev_destroy(new);
out_free_new:
	kfree(new);
	return ERR_PTR(error);
}

static void mddev_free(struct mddev *mddev)
{
	spin_lock(&all_mddevs_lock);
	list_del(&mddev->all_mddevs);
	spin_unlock(&all_mddevs_lock);

	mddev_destroy(mddev);
	kfree(mddev);
}

static const struct attribute_group md_redundancy_group;

void mddev_unlock(struct mddev *mddev)
{
	struct md_rdev *rdev;
	struct md_rdev *tmp;
	LIST_HEAD(delete);

	if (!list_empty(&mddev->deleting))
		list_splice_init(&mddev->deleting, &delete);

	if (mddev->to_remove) {
		/* These cannot be removed under reconfig_mutex as
		 * an access to the files will try to take reconfig_mutex
		 * while holding the file unremovable, which leads to
		 * a deadlock.
		 * So hold set sysfs_active while the remove in happeing,
		 * and anything else which might set ->to_remove or my
		 * otherwise change the sysfs namespace will fail with
		 * -EBUSY if sysfs_active is still set.
		 * We set sysfs_active under reconfig_mutex and elsewhere
		 * test it under the same mutex to ensure its correct value
		 * is seen.
		 */
		const struct attribute_group *to_remove = mddev->to_remove;
		mddev->to_remove = NULL;
		mddev->sysfs_active = 1;
		mutex_unlock(&mddev->reconfig_mutex);

		if (mddev->kobj.sd) {
			if (to_remove != &md_redundancy_group)
				sysfs_remove_group(&mddev->kobj, to_remove);
			if (mddev->pers == NULL ||
			    mddev->pers->sync_request == NULL) {
				sysfs_remove_group(&mddev->kobj, &md_redundancy_group);
				if (mddev->sysfs_action)
					sysfs_put(mddev->sysfs_action);
				if (mddev->sysfs_completed)
					sysfs_put(mddev->sysfs_completed);
				if (mddev->sysfs_degraded)
					sysfs_put(mddev->sysfs_degraded);
				mddev->sysfs_action = NULL;
				mddev->sysfs_completed = NULL;
				mddev->sysfs_degraded = NULL;
			}
		}
		mddev->sysfs_active = 0;
	} else
		mutex_unlock(&mddev->reconfig_mutex);

	md_wakeup_thread(mddev->thread);
	wake_up(&mddev->sb_wait);

	list_for_each_entry_safe(rdev, tmp, &delete, same_set) {
		list_del_init(&rdev->same_set);
		kobject_del(&rdev->kobj);
		export_rdev(rdev, mddev);
	}
}
EXPORT_SYMBOL_GPL(mddev_unlock);

struct md_rdev *md_find_rdev_nr_rcu(struct mddev *mddev, int nr)
{
	struct md_rdev *rdev;

	rdev_for_each_rcu(rdev, mddev)
		if (rdev->desc_nr == nr)
			return rdev;

	return NULL;
}
EXPORT_SYMBOL_GPL(md_find_rdev_nr_rcu);

static struct md_rdev *find_rdev(struct mddev *mddev, dev_t dev)
{
	struct md_rdev *rdev;

	rdev_for_each(rdev, mddev)
		if (rdev->bdev->bd_dev == dev)
			return rdev;

	return NULL;
}

struct md_rdev *md_find_rdev_rcu(struct mddev *mddev, dev_t dev)
{
	struct md_rdev *rdev;

	rdev_for_each_rcu(rdev, mddev)
		if (rdev->bdev->bd_dev == dev)
			return rdev;

	return NULL;
}
EXPORT_SYMBOL_GPL(md_find_rdev_rcu);

static struct md_personality *find_pers(int level, char *clevel)
{
	struct md_personality *pers;
	list_for_each_entry(pers, &pers_list, list) {
		if (level != LEVEL_NONE && pers->level == level)
			return pers;
		if (strcmp(pers->name, clevel)==0)
			return pers;
	}
	return NULL;
}

/* return the offset of the super block in 512byte sectors */
static inline sector_t calc_dev_sboffset(struct md_rdev *rdev)
{
	return MD_NEW_SIZE_SECTORS(bdev_nr_sectors(rdev->bdev));
}

static int alloc_disk_sb(struct md_rdev *rdev)
{
	rdev->sb_page = alloc_page(GFP_KERNEL);
	if (!rdev->sb_page)
		return -ENOMEM;
	return 0;
}

void md_rdev_clear(struct md_rdev *rdev)
{
	if (rdev->sb_page) {
		put_page(rdev->sb_page);
		rdev->sb_loaded = 0;
		rdev->sb_page = NULL;
		rdev->sb_start = 0;
		rdev->sectors = 0;
	}
	if (rdev->bb_page) {
		put_page(rdev->bb_page);
		rdev->bb_page = NULL;
	}
	badblocks_exit(&rdev->badblocks);
}
EXPORT_SYMBOL_GPL(md_rdev_clear);

static void super_written(struct bio *bio)
{
	struct md_rdev *rdev = bio->bi_private;
	struct mddev *mddev = rdev->mddev;

	if (bio->bi_status) {
		pr_err("md: %s gets error=%d\n", __func__,
		       blk_status_to_errno(bio->bi_status));
		md_error(mddev, rdev);
		if (!test_bit(Faulty, &rdev->flags)
		    && (bio->bi_opf & MD_FAILFAST)) {
			set_bit(MD_SB_NEED_REWRITE, &mddev->sb_flags);
			set_bit(LastDev, &rdev->flags);
		}
	} else
		clear_bit(LastDev, &rdev->flags);

	bio_put(bio);

	rdev_dec_pending(rdev, mddev);

	if (atomic_dec_and_test(&mddev->pending_writes))
		wake_up(&mddev->sb_wait);
}

void md_super_write(struct mddev *mddev, struct md_rdev *rdev,
		   sector_t sector, int size, struct page *page)
{
	/* write first size bytes of page to sector of rdev
	 * Increment mddev->pending_writes before returning
	 * and decrement it on completion, waking up sb_wait
	 * if zero is reached.
	 * If an error occurred, call md_error
	 */
	struct bio *bio;

	if (!page)
		return;

	if (test_bit(Faulty, &rdev->flags))
		return;

	bio = bio_alloc_bioset(rdev->meta_bdev ? rdev->meta_bdev : rdev->bdev,
			      1,
			      REQ_OP_WRITE | REQ_SYNC | REQ_IDLE | REQ_META
				  | REQ_PREFLUSH | REQ_FUA,
			      GFP_NOIO, &mddev->sync_set);

	atomic_inc(&rdev->nr_pending);

	bio->bi_iter.bi_sector = sector;
	__bio_add_page(bio, page, size, 0);
	bio->bi_private = rdev;
	bio->bi_end_io = super_written;

	if (test_bit(MD_FAILFAST_SUPPORTED, &mddev->flags) &&
	    test_bit(FailFast, &rdev->flags) &&
	    !test_bit(LastDev, &rdev->flags))
		bio->bi_opf |= MD_FAILFAST;

	atomic_inc(&mddev->pending_writes);
	submit_bio(bio);
}

int md_super_wait(struct mddev *mddev)
{
	/* wait for all superblock writes that were scheduled to complete */
	wait_event(mddev->sb_wait, atomic_read(&mddev->pending_writes)==0);
	if (test_and_clear_bit(MD_SB_NEED_REWRITE, &mddev->sb_flags))
		return -EAGAIN;
	return 0;
}

int sync_page_io(struct md_rdev *rdev, sector_t sector, int size,
		 struct page *page, blk_opf_t opf, bool metadata_op)
{
	struct bio bio;
	struct bio_vec bvec;

	if (metadata_op && rdev->meta_bdev)
		bio_init(&bio, rdev->meta_bdev, &bvec, 1, opf);
	else
		bio_init(&bio, rdev->bdev, &bvec, 1, opf);

	if (metadata_op)
		bio.bi_iter.bi_sector = sector + rdev->sb_start;
	else if (rdev->mddev->reshape_position != MaxSector &&
		 (rdev->mddev->reshape_backwards ==
		  (sector >= rdev->mddev->reshape_position)))
		bio.bi_iter.bi_sector = sector + rdev->new_data_offset;
	else
		bio.bi_iter.bi_sector = sector + rdev->data_offset;
	__bio_add_page(&bio, page, size, 0);

	submit_bio_wait(&bio);

	return !bio.bi_status;
}
EXPORT_SYMBOL_GPL(sync_page_io);

static int read_disk_sb(struct md_rdev *rdev, int size)
{
	if (rdev->sb_loaded)
		return 0;

	if (!sync_page_io(rdev, 0, size, rdev->sb_page, REQ_OP_READ, true))
		goto fail;
	rdev->sb_loaded = 1;
	return 0;

fail:
	pr_err("md: disabled device %pg, could not read superblock.\n",
	       rdev->bdev);
	return -EINVAL;
}

static int md_uuid_equal(mdp_super_t *sb1, mdp_super_t *sb2)
{
	return	sb1->set_uuid0 == sb2->set_uuid0 &&
		sb1->set_uuid1 == sb2->set_uuid1 &&
		sb1->set_uuid2 == sb2->set_uuid2 &&
		sb1->set_uuid3 == sb2->set_uuid3;
}

static int md_sb_equal(mdp_super_t *sb1, mdp_super_t *sb2)
{
	int ret;
	mdp_super_t *tmp1, *tmp2;

	tmp1 = kmalloc(sizeof(*tmp1),GFP_KERNEL);
	tmp2 = kmalloc(sizeof(*tmp2),GFP_KERNEL);

	if (!tmp1 || !tmp2) {
		ret = 0;
		goto abort;
	}

	*tmp1 = *sb1;
	*tmp2 = *sb2;

	/*
	 * nr_disks is not constant
	 */
	tmp1->nr_disks = 0;
	tmp2->nr_disks = 0;

	ret = (memcmp(tmp1, tmp2, MD_SB_GENERIC_CONSTANT_WORDS * 4) == 0);
abort:
	kfree(tmp1);
	kfree(tmp2);
	return ret;
}

static u32 md_csum_fold(u32 csum)
{
	csum = (csum & 0xffff) + (csum >> 16);
	return (csum & 0xffff) + (csum >> 16);
}

static unsigned int calc_sb_csum(mdp_super_t *sb)
{
	u64 newcsum = 0;
	u32 *sb32 = (u32*)sb;
	int i;
	unsigned int disk_csum, csum;

	disk_csum = sb->sb_csum;
	sb->sb_csum = 0;

	for (i = 0; i < MD_SB_BYTES/4 ; i++)
		newcsum += sb32[i];
	csum = (newcsum & 0xffffffff) + (newcsum>>32);

#ifdef CONFIG_ALPHA
	/* This used to use csum_partial, which was wrong for several
	 * reasons including that different results are returned on
	 * different architectures.  It isn't critical that we get exactly
	 * the same return value as before (we always csum_fold before
	 * testing, and that removes any differences).  However as we
	 * know that csum_partial always returned a 16bit value on
	 * alphas, do a fold to maximise conformity to previous behaviour.
	 */
	sb->sb_csum = md_csum_fold(disk_csum);
#else
	sb->sb_csum = disk_csum;
#endif
	return csum;
}

/*
 * Handle superblock details.
 * We want to be able to handle multiple superblock formats
 * so we have a common interface to them all, and an array of
 * different handlers.
 * We rely on user-space to write the initial superblock, and support
 * reading and updating of superblocks.
 * Interface methods are:
 *   int load_super(struct md_rdev *dev, struct md_rdev *refdev, int minor_version)
 *      loads and validates a superblock on dev.
 *      if refdev != NULL, compare superblocks on both devices
 *    Return:
 *      0 - dev has a superblock that is compatible with refdev
 *      1 - dev has a superblock that is compatible and newer than refdev
 *          so dev should be used as the refdev in future
 *     -EINVAL superblock incompatible or invalid
 *     -othererror e.g. -EIO
 *
 *   int validate_super(struct mddev *mddev, struct md_rdev *dev)
 *      Verify that dev is acceptable into mddev.
 *       The first time, mddev->raid_disks will be 0, and data from
 *       dev should be merged in.  Subsequent calls check that dev
 *       is new enough.  Return 0 or -EINVAL
 *
 *   void sync_super(struct mddev *mddev, struct md_rdev *dev)
 *     Update the superblock for rdev with data in mddev
 *     This does not write to disc.
 *
 */

struct super_type  {
	char		    *name;
	struct module	    *owner;
	int		    (*load_super)(struct md_rdev *rdev,
					  struct md_rdev *refdev,
					  int minor_version);
	int		    (*validate_super)(struct mddev *mddev,
					      struct md_rdev *freshest,
					      struct md_rdev *rdev);
	void		    (*sync_super)(struct mddev *mddev,
					  struct md_rdev *rdev);
	unsigned long long  (*rdev_size_change)(struct md_rdev *rdev,
						sector_t num_sectors);
	int		    (*allow_new_offset)(struct md_rdev *rdev,
						unsigned long long new_offset);
};

/*
 * Check that the given mddev has no bitmap.
 *
 * This function is called from the run method of all personalities that do not
 * support bitmaps. It prints an error message and returns non-zero if mddev
 * has a bitmap. Otherwise, it returns 0.
 *
 */
int md_check_no_bitmap(struct mddev *mddev)
{
	if (!mddev->bitmap_info.file && !mddev->bitmap_info.offset)
		return 0;
	pr_warn("%s: bitmaps are not supported for %s\n",
		mdname(mddev), mddev->pers->name);
	return 1;
}
EXPORT_SYMBOL(md_check_no_bitmap);

/*
 * load_super for 0.90.0
 */
static int super_90_load(struct md_rdev *rdev, struct md_rdev *refdev, int minor_version)
{
	mdp_super_t *sb;
	int ret;
	bool spare_disk = true;

	/*
	 * Calculate the position of the superblock (512byte sectors),
	 * it's at the end of the disk.
	 *
	 * It also happens to be a multiple of 4Kb.
	 */
	rdev->sb_start = calc_dev_sboffset(rdev);

	ret = read_disk_sb(rdev, MD_SB_BYTES);
	if (ret)
		return ret;

	ret = -EINVAL;

	sb = page_address(rdev->sb_page);

	if (sb->md_magic != MD_SB_MAGIC) {
		pr_warn("md: invalid raid superblock magic on %pg\n",
			rdev->bdev);
		goto abort;
	}

	if (sb->major_version != 0 ||
	    sb->minor_version < 90 ||
	    sb->minor_version > 91) {
		pr_warn("Bad version number %d.%d on %pg\n",
			sb->major_version, sb->minor_version, rdev->bdev);
		goto abort;
	}

	if (sb->raid_disks <= 0)
		goto abort;

	if (md_csum_fold(calc_sb_csum(sb)) != md_csum_fold(sb->sb_csum)) {
		pr_warn("md: invalid superblock checksum on %pg\n", rdev->bdev);
		goto abort;
	}

	rdev->preferred_minor = sb->md_minor;
	rdev->data_offset = 0;
	rdev->new_data_offset = 0;
	rdev->sb_size = MD_SB_BYTES;
	rdev->badblocks.shift = -1;

	rdev->desc_nr = sb->this_disk.number;

	/* not spare disk */
	if (rdev->desc_nr >= 0 && rdev->desc_nr < MD_SB_DISKS &&
	    sb->disks[rdev->desc_nr].state & ((1<<MD_DISK_SYNC) | (1 << MD_DISK_ACTIVE)))
		spare_disk = false;

	if (!refdev) {
		if (!spare_disk)
			ret = 1;
		else
			ret = 0;
	} else {
		__u64 ev1, ev2;
		mdp_super_t *refsb = page_address(refdev->sb_page);
		if (!md_uuid_equal(refsb, sb)) {
			pr_warn("md: %pg has different UUID to %pg\n",
				rdev->bdev, refdev->bdev);
			goto abort;
		}
		if (!md_sb_equal(refsb, sb)) {
			pr_warn("md: %pg has same UUID but different superblock to %pg\n",
				rdev->bdev, refdev->bdev);
			goto abort;
		}
		ev1 = md_event(sb);
		ev2 = md_event(refsb);

		if (!spare_disk && ev1 > ev2)
			ret = 1;
		else
			ret = 0;
	}
	rdev->sectors = rdev->sb_start;
	/* Limit to 4TB as metadata cannot record more than that.
	 * (not needed for Linear and RAID0 as metadata doesn't
	 * record this size)
	 */
	if ((u64)rdev->sectors >= (2ULL << 32) && sb->level >= 1)
		rdev->sectors = (sector_t)(2ULL << 32) - 2;

	if (rdev->sectors < ((sector_t)sb->size) * 2 && sb->level >= 1)
		/* "this cannot possibly happen" ... */
		ret = -EINVAL;

 abort:
	return ret;
}

/*
 * validate_super for 0.90.0
 * note: we are not using "freshest" for 0.9 superblock
 */
static int super_90_validate(struct mddev *mddev, struct md_rdev *freshest, struct md_rdev *rdev)
{
	mdp_disk_t *desc;
	mdp_super_t *sb = page_address(rdev->sb_page);
	__u64 ev1 = md_event(sb);

	rdev->raid_disk = -1;
	clear_bit(Faulty, &rdev->flags);
	clear_bit(In_sync, &rdev->flags);
	clear_bit(Bitmap_sync, &rdev->flags);
	clear_bit(WriteMostly, &rdev->flags);

	if (mddev->raid_disks == 0) {
		mddev->major_version = 0;
		mddev->minor_version = sb->minor_version;
		mddev->patch_version = sb->patch_version;
		mddev->external = 0;
		mddev->chunk_sectors = sb->chunk_size >> 9;
		mddev->ctime = sb->ctime;
		mddev->utime = sb->utime;
		mddev->level = sb->level;
		mddev->clevel[0] = 0;
		mddev->layout = sb->layout;
		mddev->raid_disks = sb->raid_disks;
		mddev->dev_sectors = ((sector_t)sb->size) * 2;
		mddev->events = ev1;
		mddev->bitmap_info.offset = 0;
		mddev->bitmap_info.space = 0;
		/* bitmap can use 60 K after the 4K superblocks */
		mddev->bitmap_info.default_offset = MD_SB_BYTES >> 9;
		mddev->bitmap_info.default_space = 64*2 - (MD_SB_BYTES >> 9);
		mddev->reshape_backwards = 0;

		if (mddev->minor_version >= 91) {
			mddev->reshape_position = sb->reshape_position;
			mddev->delta_disks = sb->delta_disks;
			mddev->new_level = sb->new_level;
			mddev->new_layout = sb->new_layout;
			mddev->new_chunk_sectors = sb->new_chunk >> 9;
			if (mddev->delta_disks < 0)
				mddev->reshape_backwards = 1;
		} else {
			mddev->reshape_position = MaxSector;
			mddev->delta_disks = 0;
			mddev->new_level = mddev->level;
			mddev->new_layout = mddev->layout;
			mddev->new_chunk_sectors = mddev->chunk_sectors;
		}
		if (mddev->level == 0)
			mddev->layout = -1;

		if (sb->state & (1<<MD_SB_CLEAN))
			mddev->recovery_cp = MaxSector;
		else {
			if (sb->events_hi == sb->cp_events_hi &&
				sb->events_lo == sb->cp_events_lo) {
				mddev->recovery_cp = sb->recovery_cp;
			} else
				mddev->recovery_cp = 0;
		}

		memcpy(mddev->uuid+0, &sb->set_uuid0, 4);
		memcpy(mddev->uuid+4, &sb->set_uuid1, 4);
		memcpy(mddev->uuid+8, &sb->set_uuid2, 4);
		memcpy(mddev->uuid+12,&sb->set_uuid3, 4);

		mddev->max_disks = MD_SB_DISKS;

		if (sb->state & (1<<MD_SB_BITMAP_PRESENT) &&
		    mddev->bitmap_info.file == NULL) {
			mddev->bitmap_info.offset =
				mddev->bitmap_info.default_offset;
			mddev->bitmap_info.space =
				mddev->bitmap_info.default_space;
		}

	} else if (mddev->pers == NULL) {
		/* Insist on good event counter while assembling, except
		 * for spares (which don't need an event count) */
		++ev1;
		if (sb->disks[rdev->desc_nr].state & (
			    (1<<MD_DISK_SYNC) | (1 << MD_DISK_ACTIVE)))
			if (ev1 < mddev->events)
				return -EINVAL;
	} else if (mddev->bitmap) {
		/* if adding to array with a bitmap, then we can accept an
		 * older device ... but not too old.
		 */
		if (ev1 < mddev->bitmap->events_cleared)
			return 0;
		if (ev1 < mddev->events)
			set_bit(Bitmap_sync, &rdev->flags);
	} else {
		if (ev1 < mddev->events)
			/* just a hot-add of a new device, leave raid_disk at -1 */
			return 0;
	}

	desc = sb->disks + rdev->desc_nr;

	if (desc->state & (1<<MD_DISK_FAULTY))
		set_bit(Faulty, &rdev->flags);
	else if (desc->state & (1<<MD_DISK_SYNC)) {
		set_bit(In_sync, &rdev->flags);
		rdev->raid_disk = desc->raid_disk;
		rdev->saved_raid_disk = desc->raid_disk;
	} else if (desc->state & (1<<MD_DISK_ACTIVE)) {
		/* active but not in sync implies recovery up to
		 * reshape position.  We don't know exactly where
		 * that is, so set to zero for now
		 */
		if (mddev->minor_version >= 91) {
			rdev->recovery_offset = 0;
			rdev->raid_disk = desc->raid_disk;
		}
	}
	if (desc->state & (1<<MD_DISK_WRITEMOSTLY))
		set_bit(WriteMostly, &rdev->flags);
	if (desc->state & (1<<MD_DISK_FAILFAST))
		set_bit(FailFast, &rdev->flags);
	return 0;
}

/*
 * sync_super for 0.90.0
 */
static void super_90_sync(struct mddev *mddev, struct md_rdev *rdev)
{
	mdp_super_t *sb;
	struct md_rdev *rdev2;
	int next_spare = mddev->raid_disks;

	/* make rdev->sb match mddev data..
	 *
	 * 1/ zero out disks
	 * 2/ Add info for each disk, keeping track of highest desc_nr (next_spare);
	 * 3/ any empty disks < next_spare become removed
	 *
	 * disks[0] gets initialised to REMOVED because
	 * we cannot be sure from other fields if it has
	 * been initialised or not.
	 */
	int i;
	int active=0, working=0,failed=0,spare=0,nr_disks=0;

	rdev->sb_size = MD_SB_BYTES;

	sb = page_address(rdev->sb_page);

	memset(sb, 0, sizeof(*sb));

	sb->md_magic = MD_SB_MAGIC;
	sb->major_version = mddev->major_version;
	sb->patch_version = mddev->patch_version;
	sb->gvalid_words  = 0; /* ignored */
	memcpy(&sb->set_uuid0, mddev->uuid+0, 4);
	memcpy(&sb->set_uuid1, mddev->uuid+4, 4);
	memcpy(&sb->set_uuid2, mddev->uuid+8, 4);
	memcpy(&sb->set_uuid3, mddev->uuid+12,4);

	sb->ctime = clamp_t(time64_t, mddev->ctime, 0, U32_MAX);
	sb->level = mddev->level;
	sb->size = mddev->dev_sectors / 2;
	sb->raid_disks = mddev->raid_disks;
	sb->md_minor = mddev->md_minor;
	sb->not_persistent = 0;
	sb->utime = clamp_t(time64_t, mddev->utime, 0, U32_MAX);
	sb->state = 0;
	sb->events_hi = (mddev->events>>32);
	sb->events_lo = (u32)mddev->events;

	if (mddev->reshape_position == MaxSector)
		sb->minor_version = 90;
	else {
		sb->minor_version = 91;
		sb->reshape_position = mddev->reshape_position;
		sb->new_level = mddev->new_level;
		sb->delta_disks = mddev->delta_disks;
		sb->new_layout = mddev->new_layout;
		sb->new_chunk = mddev->new_chunk_sectors << 9;
	}
	mddev->minor_version = sb->minor_version;
	if (mddev->in_sync)
	{
		sb->recovery_cp = mddev->recovery_cp;
		sb->cp_events_hi = (mddev->events>>32);
		sb->cp_events_lo = (u32)mddev->events;
		if (mddev->recovery_cp == MaxSector)
			sb->state = (1<< MD_SB_CLEAN);
	} else
		sb->recovery_cp = 0;

	sb->layout = mddev->layout;
	sb->chunk_size = mddev->chunk_sectors << 9;

	if (mddev->bitmap && mddev->bitmap_info.file == NULL)
		sb->state |= (1<<MD_SB_BITMAP_PRESENT);

	sb->disks[0].state = (1<<MD_DISK_REMOVED);
	rdev_for_each(rdev2, mddev) {
		mdp_disk_t *d;
		int desc_nr;
		int is_active = test_bit(In_sync, &rdev2->flags);

		if (rdev2->raid_disk >= 0 &&
		    sb->minor_version >= 91)
			/* we have nowhere to store the recovery_offset,
			 * but if it is not below the reshape_position,
			 * we can piggy-back on that.
			 */
			is_active = 1;
		if (rdev2->raid_disk < 0 ||
		    test_bit(Faulty, &rdev2->flags))
			is_active = 0;
		if (is_active)
			desc_nr = rdev2->raid_disk;
		else
			desc_nr = next_spare++;
		rdev2->desc_nr = desc_nr;
		d = &sb->disks[rdev2->desc_nr];
		nr_disks++;
		d->number = rdev2->desc_nr;
		d->major = MAJOR(rdev2->bdev->bd_dev);
		d->minor = MINOR(rdev2->bdev->bd_dev);
		if (is_active)
			d->raid_disk = rdev2->raid_disk;
		else
			d->raid_disk = rdev2->desc_nr; /* compatibility */
		if (test_bit(Faulty, &rdev2->flags))
			d->state = (1<<MD_DISK_FAULTY);
		else if (is_active) {
			d->state = (1<<MD_DISK_ACTIVE);
			if (test_bit(In_sync, &rdev2->flags))
				d->state |= (1<<MD_DISK_SYNC);
			active++;
			working++;
		} else {
			d->state = 0;
			spare++;
			working++;
		}
		if (test_bit(WriteMostly, &rdev2->flags))
			d->state |= (1<<MD_DISK_WRITEMOSTLY);
		if (test_bit(FailFast, &rdev2->flags))
			d->state |= (1<<MD_DISK_FAILFAST);
	}
	/* now set the "removed" and "faulty" bits on any missing devices */
	for (i=0 ; i < mddev->raid_disks ; i++) {
		mdp_disk_t *d = &sb->disks[i];
		if (d->state == 0 && d->number == 0) {
			d->number = i;
			d->raid_disk = i;
			d->state = (1<<MD_DISK_REMOVED);
			d->state |= (1<<MD_DISK_FAULTY);
			failed++;
		}
	}
	sb->nr_disks = nr_disks;
	sb->active_disks = active;
	sb->working_disks = working;
	sb->failed_disks = failed;
	sb->spare_disks = spare;

	sb->this_disk = sb->disks[rdev->desc_nr];
	sb->sb_csum = calc_sb_csum(sb);
}

/*
 * rdev_size_change for 0.90.0
 */
static unsigned long long
super_90_rdev_size_change(struct md_rdev *rdev, sector_t num_sectors)
{
	if (num_sectors && num_sectors < rdev->mddev->dev_sectors)
		return 0; /* component must fit device */
	if (rdev->mddev->bitmap_info.offset)
		return 0; /* can't move bitmap */
	rdev->sb_start = calc_dev_sboffset(rdev);
	if (!num_sectors || num_sectors > rdev->sb_start)
		num_sectors = rdev->sb_start;
	/* Limit to 4TB as metadata cannot record more than that.
	 * 4TB == 2^32 KB, or 2*2^32 sectors.
	 */
	if ((u64)num_sectors >= (2ULL << 32) && rdev->mddev->level >= 1)
		num_sectors = (sector_t)(2ULL << 32) - 2;
	do {
		md_super_write(rdev->mddev, rdev, rdev->sb_start, rdev->sb_size,
		       rdev->sb_page);
	} while (md_super_wait(rdev->mddev) < 0);
	return num_sectors;
}

static int
super_90_allow_new_offset(struct md_rdev *rdev, unsigned long long new_offset)
{
	/* non-zero offset changes not possible with v0.90 */
	return new_offset == 0;
}

/*
 * version 1 superblock
 */

static __le32 calc_sb_1_csum(struct mdp_superblock_1 *sb)
{
	__le32 disk_csum;
	u32 csum;
	unsigned long long newcsum;
	int size = 256 + le32_to_cpu(sb->max_dev)*2;
	__le32 *isuper = (__le32*)sb;

	disk_csum = sb->sb_csum;
	sb->sb_csum = 0;
	newcsum = 0;
	for (; size >= 4; size -= 4)
		newcsum += le32_to_cpu(*isuper++);

	if (size == 2)
		newcsum += le16_to_cpu(*(__le16*) isuper);

	csum = (newcsum & 0xffffffff) + (newcsum >> 32);
	sb->sb_csum = disk_csum;
	return cpu_to_le32(csum);
}

static int super_1_load(struct md_rdev *rdev, struct md_rdev *refdev, int minor_version)
{
	struct mdp_superblock_1 *sb;
	int ret;
	sector_t sb_start;
	sector_t sectors;
	int bmask;
	bool spare_disk = true;

	/*
	 * Calculate the position of the superblock in 512byte sectors.
	 * It is always aligned to a 4K boundary and
	 * depeding on minor_version, it can be:
	 * 0: At least 8K, but less than 12K, from end of device
	 * 1: At start of device
	 * 2: 4K from start of device.
	 */
	switch(minor_version) {
	case 0:
		sb_start = bdev_nr_sectors(rdev->bdev) - 8 * 2;
		sb_start &= ~(sector_t)(4*2-1);
		break;
	case 1:
		sb_start = 0;
		break;
	case 2:
		sb_start = 8;
		break;
	default:
		return -EINVAL;
	}
	rdev->sb_start = sb_start;

	/* superblock is rarely larger than 1K, but it can be larger,
	 * and it is safe to read 4k, so we do that
	 */
	ret = read_disk_sb(rdev, 4096);
	if (ret) return ret;

	sb = page_address(rdev->sb_page);

	if (sb->magic != cpu_to_le32(MD_SB_MAGIC) ||
	    sb->major_version != cpu_to_le32(1) ||
	    le32_to_cpu(sb->max_dev) > (4096-256)/2 ||
	    le64_to_cpu(sb->super_offset) != rdev->sb_start ||
	    (le32_to_cpu(sb->feature_map) & ~MD_FEATURE_ALL) != 0)
		return -EINVAL;

	if (calc_sb_1_csum(sb) != sb->sb_csum) {
		pr_warn("md: invalid superblock checksum on %pg\n",
			rdev->bdev);
		return -EINVAL;
	}
	if (le64_to_cpu(sb->data_size) < 10) {
		pr_warn("md: data_size too small on %pg\n",
			rdev->bdev);
		return -EINVAL;
	}
	if (sb->pad0 ||
	    sb->pad3[0] ||
	    memcmp(sb->pad3, sb->pad3+1, sizeof(sb->pad3) - sizeof(sb->pad3[1])))
		/* Some padding is non-zero, might be a new feature */
		return -EINVAL;

	rdev->preferred_minor = 0xffff;
	rdev->data_offset = le64_to_cpu(sb->data_offset);
	rdev->new_data_offset = rdev->data_offset;
	if ((le32_to_cpu(sb->feature_map) & MD_FEATURE_RESHAPE_ACTIVE) &&
	    (le32_to_cpu(sb->feature_map) & MD_FEATURE_NEW_OFFSET))
		rdev->new_data_offset += (s32)le32_to_cpu(sb->new_offset);
	atomic_set(&rdev->corrected_errors, le32_to_cpu(sb->cnt_corrected_read));

	rdev->sb_size = le32_to_cpu(sb->max_dev) * 2 + 256;
	bmask = queue_logical_block_size(rdev->bdev->bd_disk->queue)-1;
	if (rdev->sb_size & bmask)
		rdev->sb_size = (rdev->sb_size | bmask) + 1;

	if (minor_version
	    && rdev->data_offset < sb_start + (rdev->sb_size/512))
		return -EINVAL;
	if (minor_version
	    && rdev->new_data_offset < sb_start + (rdev->sb_size/512))
		return -EINVAL;

	rdev->desc_nr = le32_to_cpu(sb->dev_number);

	if (!rdev->bb_page) {
		rdev->bb_page = alloc_page(GFP_KERNEL);
		if (!rdev->bb_page)
			return -ENOMEM;
	}
	if ((le32_to_cpu(sb->feature_map) & MD_FEATURE_BAD_BLOCKS) &&
	    rdev->badblocks.count == 0) {
		/* need to load the bad block list.
		 * Currently we limit it to one page.
		 */
		s32 offset;
		sector_t bb_sector;
		__le64 *bbp;
		int i;
		int sectors = le16_to_cpu(sb->bblog_size);
		if (sectors > (PAGE_SIZE / 512))
			return -EINVAL;
		offset = le32_to_cpu(sb->bblog_offset);
		if (offset == 0)
			return -EINVAL;
		bb_sector = (long long)offset;
		if (!sync_page_io(rdev, bb_sector, sectors << 9,
				  rdev->bb_page, REQ_OP_READ, true))
			return -EIO;
		bbp = (__le64 *)page_address(rdev->bb_page);
		rdev->badblocks.shift = sb->bblog_shift;
		for (i = 0 ; i < (sectors << (9-3)) ; i++, bbp++) {
			u64 bb = le64_to_cpu(*bbp);
			int count = bb & (0x3ff);
			u64 sector = bb >> 10;
			sector <<= sb->bblog_shift;
			count <<= sb->bblog_shift;
			if (bb + 1 == 0)
				break;
			if (badblocks_set(&rdev->badblocks, sector, count, 1))
				return -EINVAL;
		}
	} else if (sb->bblog_offset != 0)
		rdev->badblocks.shift = 0;

	if ((le32_to_cpu(sb->feature_map) &
	    (MD_FEATURE_PPL | MD_FEATURE_MULTIPLE_PPLS))) {
		rdev->ppl.offset = (__s16)le16_to_cpu(sb->ppl.offset);
		rdev->ppl.size = le16_to_cpu(sb->ppl.size);
		rdev->ppl.sector = rdev->sb_start + rdev->ppl.offset;
	}

	if ((le32_to_cpu(sb->feature_map) & MD_FEATURE_RAID0_LAYOUT) &&
	    sb->level != 0)
		return -EINVAL;

	/* not spare disk */
	if (rdev->desc_nr >= 0 && rdev->desc_nr < le32_to_cpu(sb->max_dev) &&
	    (le16_to_cpu(sb->dev_roles[rdev->desc_nr]) < MD_DISK_ROLE_MAX ||
	     le16_to_cpu(sb->dev_roles[rdev->desc_nr]) == MD_DISK_ROLE_JOURNAL))
		spare_disk = false;

	if (!refdev) {
		if (!spare_disk)
			ret = 1;
		else
			ret = 0;
	} else {
		__u64 ev1, ev2;
		struct mdp_superblock_1 *refsb = page_address(refdev->sb_page);

		if (memcmp(sb->set_uuid, refsb->set_uuid, 16) != 0 ||
		    sb->level != refsb->level ||
		    sb->layout != refsb->layout ||
		    sb->chunksize != refsb->chunksize) {
			pr_warn("md: %pg has strangely different superblock to %pg\n",
				rdev->bdev,
				refdev->bdev);
			return -EINVAL;
		}
		ev1 = le64_to_cpu(sb->events);
		ev2 = le64_to_cpu(refsb->events);

		if (!spare_disk && ev1 > ev2)
			ret = 1;
		else
			ret = 0;
	}
	if (minor_version)
		sectors = bdev_nr_sectors(rdev->bdev) - rdev->data_offset;
	else
		sectors = rdev->sb_start;
	if (sectors < le64_to_cpu(sb->data_size))
		return -EINVAL;
	rdev->sectors = le64_to_cpu(sb->data_size);
	return ret;
}

static int super_1_validate(struct mddev *mddev, struct md_rdev *freshest, struct md_rdev *rdev)
{
	struct mdp_superblock_1 *sb = page_address(rdev->sb_page);
	__u64 ev1 = le64_to_cpu(sb->events);
	int role;

	rdev->raid_disk = -1;
	clear_bit(Faulty, &rdev->flags);
	clear_bit(In_sync, &rdev->flags);
	clear_bit(Bitmap_sync, &rdev->flags);
	clear_bit(WriteMostly, &rdev->flags);

	if (mddev->raid_disks == 0) {
		mddev->major_version = 1;
		mddev->patch_version = 0;
		mddev->external = 0;
		mddev->chunk_sectors = le32_to_cpu(sb->chunksize);
		mddev->ctime = le64_to_cpu(sb->ctime);
		mddev->utime = le64_to_cpu(sb->utime);
		mddev->level = le32_to_cpu(sb->level);
		mddev->clevel[0] = 0;
		mddev->layout = le32_to_cpu(sb->layout);
		mddev->raid_disks = le32_to_cpu(sb->raid_disks);
		mddev->dev_sectors = le64_to_cpu(sb->size);
		mddev->events = ev1;
		mddev->bitmap_info.offset = 0;
		mddev->bitmap_info.space = 0;
		/* Default location for bitmap is 1K after superblock
		 * using 3K - total of 4K
		 */
		mddev->bitmap_info.default_offset = 1024 >> 9;
		mddev->bitmap_info.default_space = (4096-1024) >> 9;
		mddev->reshape_backwards = 0;

		mddev->recovery_cp = le64_to_cpu(sb->resync_offset);
		memcpy(mddev->uuid, sb->set_uuid, 16);

		mddev->max_disks =  (4096-256)/2;

		if ((le32_to_cpu(sb->feature_map) & MD_FEATURE_BITMAP_OFFSET) &&
		    mddev->bitmap_info.file == NULL) {
			mddev->bitmap_info.offset =
				(__s32)le32_to_cpu(sb->bitmap_offset);
			/* Metadata doesn't record how much space is available.
			 * For 1.0, we assume we can use up to the superblock
			 * if before, else to 4K beyond superblock.
			 * For others, assume no change is possible.
			 */
			if (mddev->minor_version > 0)
				mddev->bitmap_info.space = 0;
			else if (mddev->bitmap_info.offset > 0)
				mddev->bitmap_info.space =
					8 - mddev->bitmap_info.offset;
			else
				mddev->bitmap_info.space =
					-mddev->bitmap_info.offset;
		}

		if ((le32_to_cpu(sb->feature_map) & MD_FEATURE_RESHAPE_ACTIVE)) {
			mddev->reshape_position = le64_to_cpu(sb->reshape_position);
			mddev->delta_disks = le32_to_cpu(sb->delta_disks);
			mddev->new_level = le32_to_cpu(sb->new_level);
			mddev->new_layout = le32_to_cpu(sb->new_layout);
			mddev->new_chunk_sectors = le32_to_cpu(sb->new_chunk);
			if (mddev->delta_disks < 0 ||
			    (mddev->delta_disks == 0 &&
			     (le32_to_cpu(sb->feature_map)
			      & MD_FEATURE_RESHAPE_BACKWARDS)))
				mddev->reshape_backwards = 1;
		} else {
			mddev->reshape_position = MaxSector;
			mddev->delta_disks = 0;
			mddev->new_level = mddev->level;
			mddev->new_layout = mddev->layout;
			mddev->new_chunk_sectors = mddev->chunk_sectors;
		}

		if (mddev->level == 0 &&
		    !(le32_to_cpu(sb->feature_map) & MD_FEATURE_RAID0_LAYOUT))
			mddev->layout = -1;

		if (le32_to_cpu(sb->feature_map) & MD_FEATURE_JOURNAL)
			set_bit(MD_HAS_JOURNAL, &mddev->flags);

		if (le32_to_cpu(sb->feature_map) &
		    (MD_FEATURE_PPL | MD_FEATURE_MULTIPLE_PPLS)) {
			if (le32_to_cpu(sb->feature_map) &
			    (MD_FEATURE_BITMAP_OFFSET | MD_FEATURE_JOURNAL))
				return -EINVAL;
			if ((le32_to_cpu(sb->feature_map) & MD_FEATURE_PPL) &&
			    (le32_to_cpu(sb->feature_map) &
					    MD_FEATURE_MULTIPLE_PPLS))
				return -EINVAL;
			set_bit(MD_HAS_PPL, &mddev->flags);
		}
	} else if (mddev->pers == NULL) {
		/* Insist of good event counter while assembling, except for
		 * spares (which don't need an event count).
		 * Similar to mdadm, we allow event counter difference of 1
		 * from the freshest device.
		 */
		if (rdev->desc_nr >= 0 &&
		    rdev->desc_nr < le32_to_cpu(sb->max_dev) &&
		    (le16_to_cpu(sb->dev_roles[rdev->desc_nr]) < MD_DISK_ROLE_MAX ||
		     le16_to_cpu(sb->dev_roles[rdev->desc_nr]) == MD_DISK_ROLE_JOURNAL))
			if (ev1 + 1 < mddev->events)
				return -EINVAL;
	} else if (mddev->bitmap) {
		/* If adding to array with a bitmap, then we can accept an
		 * older device, but not too old.
		 */
		if (ev1 < mddev->bitmap->events_cleared)
			return 0;
		if (ev1 < mddev->events)
			set_bit(Bitmap_sync, &rdev->flags);
	} else {
		if (ev1 < mddev->events)
			/* just a hot-add of a new device, leave raid_disk at -1 */
			return 0;
	}

	if (rdev->desc_nr < 0 ||
	    rdev->desc_nr >= le32_to_cpu(sb->max_dev)) {
		role = MD_DISK_ROLE_SPARE;
		rdev->desc_nr = -1;
	} else if (mddev->pers == NULL && freshest && ev1 < mddev->events) {
		/*
		 * If we are assembling, and our event counter is smaller than the
		 * highest event counter, we cannot trust our superblock about the role.
		 * It could happen that our rdev was marked as Faulty, and all other
		 * superblocks were updated with +1 event counter.
		 * Then, before the next superblock update, which typically happens when
		 * remove_and_add_spares() removes the device from the array, there was
		 * a crash or reboot.
		 * If we allow current rdev without consulting the freshest superblock,
		 * we could cause data corruption.
		 * Note that in this case our event counter is smaller by 1 than the
		 * highest, otherwise, this rdev would not be allowed into array;
		 * both kernel and mdadm allow event counter difference of 1.
		 */
		struct mdp_superblock_1 *freshest_sb = page_address(freshest->sb_page);
		u32 freshest_max_dev = le32_to_cpu(freshest_sb->max_dev);

		if (rdev->desc_nr >= freshest_max_dev) {
			/* this is unexpected, better not proceed */
			pr_warn("md: %s: rdev[%pg]: desc_nr(%d) >= freshest(%pg)->sb->max_dev(%u)\n",
				mdname(mddev), rdev->bdev, rdev->desc_nr,
				freshest->bdev, freshest_max_dev);
			return -EUCLEAN;
		}

		role = le16_to_cpu(freshest_sb->dev_roles[rdev->desc_nr]);
		pr_debug("md: %s: rdev[%pg]: role=%d(0x%x) according to freshest %pg\n",
			 mdname(mddev), rdev->bdev, role, role, freshest->bdev);
	} else {
		role = le16_to_cpu(sb->dev_roles[rdev->desc_nr]);
	}
	switch (role) {
	case MD_DISK_ROLE_SPARE: /* spare */
		break;
	case MD_DISK_ROLE_FAULTY: /* faulty */
		set_bit(Faulty, &rdev->flags);
		break;
	case MD_DISK_ROLE_JOURNAL: /* journal device */
		if (!(le32_to_cpu(sb->feature_map) & MD_FEATURE_JOURNAL)) {
			/* journal device without journal feature */
			pr_warn("md: journal device provided without journal feature, ignoring the device\n");
			return -EINVAL;
		}
		set_bit(Journal, &rdev->flags);
		rdev->journal_tail = le64_to_cpu(sb->journal_tail);
		rdev->raid_disk = 0;
		break;
	default:
		rdev->saved_raid_disk = role;
		if ((le32_to_cpu(sb->feature_map) &
		     MD_FEATURE_RECOVERY_OFFSET)) {
			rdev->recovery_offset = le64_to_cpu(sb->recovery_offset);
			if (!(le32_to_cpu(sb->feature_map) &
			      MD_FEATURE_RECOVERY_BITMAP))
				rdev->saved_raid_disk = -1;
		} else {
			/*
			 * If the array is FROZEN, then the device can't
			 * be in_sync with rest of array.
			 */
			if (!test_bit(MD_RECOVERY_FROZEN,
				      &mddev->recovery))
				set_bit(In_sync, &rdev->flags);
		}
		rdev->raid_disk = role;
		break;
	}
	if (sb->devflags & WriteMostly1)
		set_bit(WriteMostly, &rdev->flags);
	if (sb->devflags & FailFast1)
		set_bit(FailFast, &rdev->flags);
	if (le32_to_cpu(sb->feature_map) & MD_FEATURE_REPLACEMENT)
		set_bit(Replacement, &rdev->flags);

	return 0;
}

static void super_1_sync(struct mddev *mddev, struct md_rdev *rdev)
{
	struct mdp_superblock_1 *sb;
	struct md_rdev *rdev2;
	int max_dev, i;
	/* make rdev->sb match mddev and rdev data. */

	sb = page_address(rdev->sb_page);

	sb->feature_map = 0;
	sb->pad0 = 0;
	sb->recovery_offset = cpu_to_le64(0);
	memset(sb->pad3, 0, sizeof(sb->pad3));

	sb->utime = cpu_to_le64((__u64)mddev->utime);
	sb->events = cpu_to_le64(mddev->events);
	if (mddev->in_sync)
		sb->resync_offset = cpu_to_le64(mddev->recovery_cp);
	else if (test_bit(MD_JOURNAL_CLEAN, &mddev->flags))
		sb->resync_offset = cpu_to_le64(MaxSector);
	else
		sb->resync_offset = cpu_to_le64(0);

	sb->cnt_corrected_read = cpu_to_le32(atomic_read(&rdev->corrected_errors));

	sb->raid_disks = cpu_to_le32(mddev->raid_disks);
	sb->size = cpu_to_le64(mddev->dev_sectors);
	sb->chunksize = cpu_to_le32(mddev->chunk_sectors);
	sb->level = cpu_to_le32(mddev->level);
	sb->layout = cpu_to_le32(mddev->layout);
	if (test_bit(FailFast, &rdev->flags))
		sb->devflags |= FailFast1;
	else
		sb->devflags &= ~FailFast1;

	if (test_bit(WriteMostly, &rdev->flags))
		sb->devflags |= WriteMostly1;
	else
		sb->devflags &= ~WriteMostly1;
	sb->data_offset = cpu_to_le64(rdev->data_offset);
	sb->data_size = cpu_to_le64(rdev->sectors);

	if (mddev->bitmap && mddev->bitmap_info.file == NULL) {
		sb->bitmap_offset = cpu_to_le32((__u32)mddev->bitmap_info.offset);
		sb->feature_map = cpu_to_le32(MD_FEATURE_BITMAP_OFFSET);
	}

	if (rdev->raid_disk >= 0 && !test_bit(Journal, &rdev->flags) &&
	    !test_bit(In_sync, &rdev->flags)) {
		sb->feature_map |=
			cpu_to_le32(MD_FEATURE_RECOVERY_OFFSET);
		sb->recovery_offset =
			cpu_to_le64(rdev->recovery_offset);
		if (rdev->saved_raid_disk >= 0 && mddev->bitmap)
			sb->feature_map |=
				cpu_to_le32(MD_FEATURE_RECOVERY_BITMAP);
	}
	/* Note: recovery_offset and journal_tail share space  */
	if (test_bit(Journal, &rdev->flags))
		sb->journal_tail = cpu_to_le64(rdev->journal_tail);
	if (test_bit(Replacement, &rdev->flags))
		sb->feature_map |=
			cpu_to_le32(MD_FEATURE_REPLACEMENT);

	if (mddev->reshape_position != MaxSector) {
		sb->feature_map |= cpu_to_le32(MD_FEATURE_RESHAPE_ACTIVE);
		sb->reshape_position = cpu_to_le64(mddev->reshape_position);
		sb->new_layout = cpu_to_le32(mddev->new_layout);
		sb->delta_disks = cpu_to_le32(mddev->delta_disks);
		sb->new_level = cpu_to_le32(mddev->new_level);
		sb->new_chunk = cpu_to_le32(mddev->new_chunk_sectors);
		if (mddev->delta_disks == 0 &&
		    mddev->reshape_backwards)
			sb->feature_map
				|= cpu_to_le32(MD_FEATURE_RESHAPE_BACKWARDS);
		if (rdev->new_data_offset != rdev->data_offset) {
			sb->feature_map
				|= cpu_to_le32(MD_FEATURE_NEW_OFFSET);
			sb->new_offset = cpu_to_le32((__u32)(rdev->new_data_offset
							     - rdev->data_offset));
		}
	}

	if (mddev_is_clustered(mddev))
		sb->feature_map |= cpu_to_le32(MD_FEATURE_CLUSTERED);

	if (rdev->badblocks.count == 0)
		/* Nothing to do for bad blocks*/ ;
	else if (sb->bblog_offset == 0)
		/* Cannot record bad blocks on this device */
		md_error(mddev, rdev);
	else {
		struct badblocks *bb = &rdev->badblocks;
		__le64 *bbp = (__le64 *)page_address(rdev->bb_page);
		u64 *p = bb->page;
		sb->feature_map |= cpu_to_le32(MD_FEATURE_BAD_BLOCKS);
		if (bb->changed) {
			unsigned seq;

retry:
			seq = read_seqbegin(&bb->lock);

			memset(bbp, 0xff, PAGE_SIZE);

			for (i = 0 ; i < bb->count ; i++) {
				u64 internal_bb = p[i];
				u64 store_bb = ((BB_OFFSET(internal_bb) << 10)
						| BB_LEN(internal_bb));
				bbp[i] = cpu_to_le64(store_bb);
			}
			bb->changed = 0;
			if (read_seqretry(&bb->lock, seq))
				goto retry;

			bb->sector = (rdev->sb_start +
				      (int)le32_to_cpu(sb->bblog_offset));
			bb->size = le16_to_cpu(sb->bblog_size);
		}
	}

	max_dev = 0;
	rdev_for_each(rdev2, mddev)
		if (rdev2->desc_nr+1 > max_dev)
			max_dev = rdev2->desc_nr+1;

	if (max_dev > le32_to_cpu(sb->max_dev)) {
		int bmask;
		sb->max_dev = cpu_to_le32(max_dev);
		rdev->sb_size = max_dev * 2 + 256;
		bmask = queue_logical_block_size(rdev->bdev->bd_disk->queue)-1;
		if (rdev->sb_size & bmask)
			rdev->sb_size = (rdev->sb_size | bmask) + 1;
	} else
		max_dev = le32_to_cpu(sb->max_dev);

	for (i=0; i<max_dev;i++)
		sb->dev_roles[i] = cpu_to_le16(MD_DISK_ROLE_SPARE);

	if (test_bit(MD_HAS_JOURNAL, &mddev->flags))
		sb->feature_map |= cpu_to_le32(MD_FEATURE_JOURNAL);

	if (test_bit(MD_HAS_PPL, &mddev->flags)) {
		if (test_bit(MD_HAS_MULTIPLE_PPLS, &mddev->flags))
			sb->feature_map |=
			    cpu_to_le32(MD_FEATURE_MULTIPLE_PPLS);
		else
			sb->feature_map |= cpu_to_le32(MD_FEATURE_PPL);
		sb->ppl.offset = cpu_to_le16(rdev->ppl.offset);
		sb->ppl.size = cpu_to_le16(rdev->ppl.size);
	}

	rdev_for_each(rdev2, mddev) {
		i = rdev2->desc_nr;
		if (test_bit(Faulty, &rdev2->flags))
			sb->dev_roles[i] = cpu_to_le16(MD_DISK_ROLE_FAULTY);
		else if (test_bit(In_sync, &rdev2->flags))
			sb->dev_roles[i] = cpu_to_le16(rdev2->raid_disk);
		else if (test_bit(Journal, &rdev2->flags))
			sb->dev_roles[i] = cpu_to_le16(MD_DISK_ROLE_JOURNAL);
		else if (rdev2->raid_disk >= 0)
			sb->dev_roles[i] = cpu_to_le16(rdev2->raid_disk);
		else
			sb->dev_roles[i] = cpu_to_le16(MD_DISK_ROLE_SPARE);
	}

	sb->sb_csum = calc_sb_1_csum(sb);
}

static sector_t super_1_choose_bm_space(sector_t dev_size)
{
	sector_t bm_space;

	/* if the device is bigger than 8Gig, save 64k for bitmap
	 * usage, if bigger than 200Gig, save 128k
	 */
	if (dev_size < 64*2)
		bm_space = 0;
	else if (dev_size - 64*2 >= 200*1024*1024*2)
		bm_space = 128*2;
	else if (dev_size - 4*2 > 8*1024*1024*2)
		bm_space = 64*2;
	else
		bm_space = 4*2;
	return bm_space;
}

static unsigned long long
super_1_rdev_size_change(struct md_rdev *rdev, sector_t num_sectors)
{
	struct mdp_superblock_1 *sb;
	sector_t max_sectors;
	if (num_sectors && num_sectors < rdev->mddev->dev_sectors)
		return 0; /* component must fit device */
	if (rdev->data_offset != rdev->new_data_offset)
		return 0; /* too confusing */
	if (rdev->sb_start < rdev->data_offset) {
		/* minor versions 1 and 2; superblock before data */
		max_sectors = bdev_nr_sectors(rdev->bdev) - rdev->data_offset;
		if (!num_sectors || num_sectors > max_sectors)
			num_sectors = max_sectors;
	} else if (rdev->mddev->bitmap_info.offset) {
		/* minor version 0 with bitmap we can't move */
		return 0;
	} else {
		/* minor version 0; superblock after data */
		sector_t sb_start, bm_space;
		sector_t dev_size = bdev_nr_sectors(rdev->bdev);

		/* 8K is for superblock */
		sb_start = dev_size - 8*2;
		sb_start &= ~(sector_t)(4*2 - 1);

		bm_space = super_1_choose_bm_space(dev_size);

		/* Space that can be used to store date needs to decrease
		 * superblock bitmap space and bad block space(4K)
		 */
		max_sectors = sb_start - bm_space - 4*2;

		if (!num_sectors || num_sectors > max_sectors)
			num_sectors = max_sectors;
		rdev->sb_start = sb_start;
	}
	sb = page_address(rdev->sb_page);
	sb->data_size = cpu_to_le64(num_sectors);
	sb->super_offset = cpu_to_le64(rdev->sb_start);
	sb->sb_csum = calc_sb_1_csum(sb);
	do {
		md_super_write(rdev->mddev, rdev, rdev->sb_start, rdev->sb_size,
			       rdev->sb_page);
	} while (md_super_wait(rdev->mddev) < 0);
	return num_sectors;

}

static int
super_1_allow_new_offset(struct md_rdev *rdev,
			 unsigned long long new_offset)
{
	/* All necessary checks on new >= old have been done */
	struct bitmap *bitmap;
	if (new_offset >= rdev->data_offset)
		return 1;

	/* with 1.0 metadata, there is no metadata to tread on
	 * so we can always move back */
	if (rdev->mddev->minor_version == 0)
		return 1;

	/* otherwise we must be sure not to step on
	 * any metadata, so stay:
	 * 36K beyond start of superblock
	 * beyond end of badblocks
	 * beyond write-intent bitmap
	 */
	if (rdev->sb_start + (32+4)*2 > new_offset)
		return 0;
	bitmap = rdev->mddev->bitmap;
	if (bitmap && !rdev->mddev->bitmap_info.file &&
	    rdev->sb_start + rdev->mddev->bitmap_info.offset +
	    bitmap->storage.file_pages * (PAGE_SIZE>>9) > new_offset)
		return 0;
	if (rdev->badblocks.sector + rdev->badblocks.size > new_offset)
		return 0;

	return 1;
}

static struct super_type super_types[] = {
	[0] = {
		.name	= "0.90.0",
		.owner	= THIS_MODULE,
		.load_super	    = super_90_load,
		.validate_super	    = super_90_validate,
		.sync_super	    = super_90_sync,
		.rdev_size_change   = super_90_rdev_size_change,
		.allow_new_offset   = super_90_allow_new_offset,
	},
	[1] = {
		.name	= "md-1",
		.owner	= THIS_MODULE,
		.load_super	    = super_1_load,
		.validate_super	    = super_1_validate,
		.sync_super	    = super_1_sync,
		.rdev_size_change   = super_1_rdev_size_change,
		.allow_new_offset   = super_1_allow_new_offset,
	},
};

static void sync_super(struct mddev *mddev, struct md_rdev *rdev)
{
	if (mddev->sync_super) {
		mddev->sync_super(mddev, rdev);
		return;
	}

	BUG_ON(mddev->major_version >= ARRAY_SIZE(super_types));

	super_types[mddev->major_version].sync_super(mddev, rdev);
}

static int match_mddev_units(struct mddev *mddev1, struct mddev *mddev2)
{
	struct md_rdev *rdev, *rdev2;

	rcu_read_lock();
	rdev_for_each_rcu(rdev, mddev1) {
		if (test_bit(Faulty, &rdev->flags) ||
		    test_bit(Journal, &rdev->flags) ||
		    rdev->raid_disk == -1)
			continue;
		rdev_for_each_rcu(rdev2, mddev2) {
			if (test_bit(Faulty, &rdev2->flags) ||
			    test_bit(Journal, &rdev2->flags) ||
			    rdev2->raid_disk == -1)
				continue;
			if (rdev->bdev->bd_disk == rdev2->bdev->bd_disk) {
				rcu_read_unlock();
				return 1;
			}
		}
	}
	rcu_read_unlock();
	return 0;
}

static LIST_HEAD(pending_raid_disks);

/*
 * Try to register data integrity profile for an mddev
 *
 * This is called when an array is started and after a disk has been kicked
 * from the array. It only succeeds if all working and active component devices
 * are integrity capable with matching profiles.
 */
int md_integrity_register(struct mddev *mddev)
{
	if (list_empty(&mddev->disks))
		return 0; /* nothing to do */
	if (mddev_is_dm(mddev) || !blk_get_integrity(mddev->gendisk))
		return 0; /* shouldn't register */

	pr_debug("md: data integrity enabled on %s\n", mdname(mddev));
	if (bioset_integrity_create(&mddev->bio_set, BIO_POOL_SIZE) ||
	    (mddev->level != 1 && mddev->level != 10 &&
	     bioset_integrity_create(&mddev->io_clone_set, BIO_POOL_SIZE))) {
		/*
		 * No need to handle the failure of bioset_integrity_create,
		 * because the function is called by md_run() -> pers->run(),
		 * md_run calls bioset_exit -> bioset_integrity_free in case
		 * of failure case.
		 */
		pr_err("md: failed to create integrity pool for %s\n",
		       mdname(mddev));
		return -EINVAL;
	}
	return 0;
}
EXPORT_SYMBOL(md_integrity_register);

static bool rdev_read_only(struct md_rdev *rdev)
{
	return bdev_read_only(rdev->bdev) ||
		(rdev->meta_bdev && bdev_read_only(rdev->meta_bdev));
}

static int bind_rdev_to_array(struct md_rdev *rdev, struct mddev *mddev)
{
	char b[BDEVNAME_SIZE];
	int err;

	/* prevent duplicates */
	if (find_rdev(mddev, rdev->bdev->bd_dev))
		return -EEXIST;

	if (rdev_read_only(rdev) && mddev->pers)
		return -EROFS;

	/* make sure rdev->sectors exceeds mddev->dev_sectors */
	if (!test_bit(Journal, &rdev->flags) &&
	    rdev->sectors &&
	    (mddev->dev_sectors == 0 || rdev->sectors < mddev->dev_sectors)) {
		if (mddev->pers) {
			/* Cannot change size, so fail
			 * If mddev->level <= 0, then we don't care
			 * about aligning sizes (e.g. linear)
			 */
			if (mddev->level > 0)
				return -ENOSPC;
		} else
			mddev->dev_sectors = rdev->sectors;
	}

	/* Verify rdev->desc_nr is unique.
	 * If it is -1, assign a free number, else
	 * check number is not in use
	 */
	rcu_read_lock();
	if (rdev->desc_nr < 0) {
		int choice = 0;
		if (mddev->pers)
			choice = mddev->raid_disks;
		while (md_find_rdev_nr_rcu(mddev, choice))
			choice++;
		rdev->desc_nr = choice;
	} else {
		if (md_find_rdev_nr_rcu(mddev, rdev->desc_nr)) {
			rcu_read_unlock();
			return -EBUSY;
		}
	}
	rcu_read_unlock();
	if (!test_bit(Journal, &rdev->flags) &&
	    mddev->max_disks && rdev->desc_nr >= mddev->max_disks) {
		pr_warn("md: %s: array is limited to %d devices\n",
			mdname(mddev), mddev->max_disks);
		return -EBUSY;
	}
	snprintf(b, sizeof(b), "%pg", rdev->bdev);
	strreplace(b, '/', '!');

	rdev->mddev = mddev;
	pr_debug("md: bind<%s>\n", b);

	if (mddev->raid_disks)
		mddev_create_serial_pool(mddev, rdev);

	if ((err = kobject_add(&rdev->kobj, &mddev->kobj, "dev-%s", b)))
		goto fail;

	/* failure here is OK */
	err = sysfs_create_link(&rdev->kobj, bdev_kobj(rdev->bdev), "block");
	rdev->sysfs_state = sysfs_get_dirent_safe(rdev->kobj.sd, "state");
	rdev->sysfs_unack_badblocks =
		sysfs_get_dirent_safe(rdev->kobj.sd, "unacknowledged_bad_blocks");
	rdev->sysfs_badblocks =
		sysfs_get_dirent_safe(rdev->kobj.sd, "bad_blocks");

	list_add_rcu(&rdev->same_set, &mddev->disks);
	bd_link_disk_holder(rdev->bdev, mddev->gendisk);

	/* May as well allow recovery to be retried once */
	mddev->recovery_disabled++;

	return 0;

 fail:
	pr_warn("md: failed to register dev-%s for %s\n",
		b, mdname(mddev));
	mddev_destroy_serial_pool(mddev, rdev);
	return err;
}

void md_autodetect_dev(dev_t dev);

/* just for claiming the bdev */
static struct md_rdev claim_rdev;

static void export_rdev(struct md_rdev *rdev, struct mddev *mddev)
{
	pr_debug("md: export_rdev(%pg)\n", rdev->bdev);
	md_rdev_clear(rdev);
#ifndef MODULE
	if (test_bit(AutoDetected, &rdev->flags))
		md_autodetect_dev(rdev->bdev->bd_dev);
#endif
	fput(rdev->bdev_file);
	rdev->bdev = NULL;
	kobject_put(&rdev->kobj);
}

static void md_kick_rdev_from_array(struct md_rdev *rdev)
{
	struct mddev *mddev = rdev->mddev;

	bd_unlink_disk_holder(rdev->bdev, rdev->mddev->gendisk);
	list_del_rcu(&rdev->same_set);
	pr_debug("md: unbind<%pg>\n", rdev->bdev);
	mddev_destroy_serial_pool(rdev->mddev, rdev);
	WRITE_ONCE(rdev->mddev, NULL);
	sysfs_remove_link(&rdev->kobj, "block");
	sysfs_put(rdev->sysfs_state);
	sysfs_put(rdev->sysfs_unack_badblocks);
	sysfs_put(rdev->sysfs_badblocks);
	rdev->sysfs_state = NULL;
	rdev->sysfs_unack_badblocks = NULL;
	rdev->sysfs_badblocks = NULL;
	rdev->badblocks.count = 0;

	synchronize_rcu();

	/*
	 * kobject_del() will wait for all in progress writers to be done, where
	 * reconfig_mutex is held, hence it can't be called under
	 * reconfig_mutex and it's delayed to mddev_unlock().
	 */
	list_add(&rdev->same_set, &mddev->deleting);
}

static void export_array(struct mddev *mddev)
{
	struct md_rdev *rdev;

	while (!list_empty(&mddev->disks)) {
		rdev = list_first_entry(&mddev->disks, struct md_rdev,
					same_set);
		md_kick_rdev_from_array(rdev);
	}
	mddev->raid_disks = 0;
	mddev->major_version = 0;
}

static bool set_in_sync(struct mddev *mddev)
{
	lockdep_assert_held(&mddev->lock);
	if (!mddev->in_sync) {
		mddev->sync_checkers++;
		spin_unlock(&mddev->lock);
		percpu_ref_switch_to_atomic_sync(&mddev->writes_pending);
		spin_lock(&mddev->lock);
		if (!mddev->in_sync &&
		    percpu_ref_is_zero(&mddev->writes_pending)) {
			mddev->in_sync = 1;
			/*
			 * Ensure ->in_sync is visible before we clear
			 * ->sync_checkers.
			 */
			smp_mb();
			set_bit(MD_SB_CHANGE_CLEAN, &mddev->sb_flags);
			sysfs_notify_dirent_safe(mddev->sysfs_state);
		}
		if (--mddev->sync_checkers == 0)
			percpu_ref_switch_to_percpu(&mddev->writes_pending);
	}
	if (mddev->safemode == 1)
		mddev->safemode = 0;
	return mddev->in_sync;
}

static void sync_sbs(struct mddev *mddev, int nospares)
{
	/* Update each superblock (in-memory image), but
	 * if we are allowed to, skip spares which already
	 * have the right event counter, or have one earlier
	 * (which would mean they aren't being marked as dirty
	 * with the rest of the array)
	 */
	struct md_rdev *rdev;
	rdev_for_each(rdev, mddev) {
		if (rdev->sb_events == mddev->events ||
		    (nospares &&
		     rdev->raid_disk < 0 &&
		     rdev->sb_events+1 == mddev->events)) {
			/* Don't update this superblock */
			rdev->sb_loaded = 2;
		} else {
			sync_super(mddev, rdev);
			rdev->sb_loaded = 1;
		}
	}
}

static bool does_sb_need_changing(struct mddev *mddev)
{
	struct md_rdev *rdev = NULL, *iter;
	struct mdp_superblock_1 *sb;
	int role;

	/* Find a good rdev */
	rdev_for_each(iter, mddev)
		if ((iter->raid_disk >= 0) && !test_bit(Faulty, &iter->flags)) {
			rdev = iter;
			break;
		}

	/* No good device found. */
	if (!rdev)
		return false;

	sb = page_address(rdev->sb_page);
	/* Check if a device has become faulty or a spare become active */
	rdev_for_each(rdev, mddev) {
		role = le16_to_cpu(sb->dev_roles[rdev->desc_nr]);
		/* Device activated? */
		if (role == MD_DISK_ROLE_SPARE && rdev->raid_disk >= 0 &&
		    !test_bit(Faulty, &rdev->flags))
			return true;
		/* Device turned faulty? */
		if (test_bit(Faulty, &rdev->flags) && (role < MD_DISK_ROLE_MAX))
			return true;
	}

	/* Check if any mddev parameters have changed */
	if ((mddev->dev_sectors != le64_to_cpu(sb->size)) ||
	    (mddev->reshape_position != le64_to_cpu(sb->reshape_position)) ||
	    (mddev->layout != le32_to_cpu(sb->layout)) ||
	    (mddev->raid_disks != le32_to_cpu(sb->raid_disks)) ||
	    (mddev->chunk_sectors != le32_to_cpu(sb->chunksize)))
		return true;

	return false;
}

void md_update_sb(struct mddev *mddev, int force_change)
{
	struct md_rdev *rdev;
	int sync_req;
	int nospares = 0;
	int any_badblocks_changed = 0;
	int ret = -1;

	if (!md_is_rdwr(mddev)) {
		if (force_change)
			set_bit(MD_SB_CHANGE_DEVS, &mddev->sb_flags);
		return;
	}

repeat:
	if (mddev_is_clustered(mddev)) {
		if (test_and_clear_bit(MD_SB_CHANGE_DEVS, &mddev->sb_flags))
			force_change = 1;
		if (test_and_clear_bit(MD_SB_CHANGE_CLEAN, &mddev->sb_flags))
			nospares = 1;
		ret = md_cluster_ops->metadata_update_start(mddev);
		/* Has someone else has updated the sb */
		if (!does_sb_need_changing(mddev)) {
			if (ret == 0)
				md_cluster_ops->metadata_update_cancel(mddev);
			bit_clear_unless(&mddev->sb_flags, BIT(MD_SB_CHANGE_PENDING),
							 BIT(MD_SB_CHANGE_DEVS) |
							 BIT(MD_SB_CHANGE_CLEAN));
			return;
		}
	}

	/*
	 * First make sure individual recovery_offsets are correct
	 * curr_resync_completed can only be used during recovery.
	 * During reshape/resync it might use array-addresses rather
	 * that device addresses.
	 */
	rdev_for_each(rdev, mddev) {
		if (rdev->raid_disk >= 0 &&
		    mddev->delta_disks >= 0 &&
		    test_bit(MD_RECOVERY_RUNNING, &mddev->recovery) &&
		    test_bit(MD_RECOVERY_RECOVER, &mddev->recovery) &&
		    !test_bit(MD_RECOVERY_RESHAPE, &mddev->recovery) &&
		    !test_bit(Journal, &rdev->flags) &&
		    !test_bit(In_sync, &rdev->flags) &&
		    mddev->curr_resync_completed > rdev->recovery_offset)
				rdev->recovery_offset = mddev->curr_resync_completed;

	}
	if (!mddev->persistent) {
		clear_bit(MD_SB_CHANGE_CLEAN, &mddev->sb_flags);
		clear_bit(MD_SB_CHANGE_DEVS, &mddev->sb_flags);
		if (!mddev->external) {
			clear_bit(MD_SB_CHANGE_PENDING, &mddev->sb_flags);
			rdev_for_each(rdev, mddev) {
				if (rdev->badblocks.changed) {
					rdev->badblocks.changed = 0;
					ack_all_badblocks(&rdev->badblocks);
					md_error(mddev, rdev);
				}
				clear_bit(Blocked, &rdev->flags);
				clear_bit(BlockedBadBlocks, &rdev->flags);
				wake_up(&rdev->blocked_wait);
			}
		}
		wake_up(&mddev->sb_wait);
		return;
	}

	spin_lock(&mddev->lock);

	mddev->utime = ktime_get_real_seconds();

	if (test_and_clear_bit(MD_SB_CHANGE_DEVS, &mddev->sb_flags))
		force_change = 1;
	if (test_and_clear_bit(MD_SB_CHANGE_CLEAN, &mddev->sb_flags))
		/* just a clean<-> dirty transition, possibly leave spares alone,
		 * though if events isn't the right even/odd, we will have to do
		 * spares after all
		 */
		nospares = 1;
	if (force_change)
		nospares = 0;
	if (mddev->degraded)
		/* If the array is degraded, then skipping spares is both
		 * dangerous and fairly pointless.
		 * Dangerous because a device that was removed from the array
		 * might have a event_count that still looks up-to-date,
		 * so it can be re-added without a resync.
		 * Pointless because if there are any spares to skip,
		 * then a recovery will happen and soon that array won't
		 * be degraded any more and the spare can go back to sleep then.
		 */
		nospares = 0;

	sync_req = mddev->in_sync;

	/* If this is just a dirty<->clean transition, and the array is clean
	 * and 'events' is odd, we can roll back to the previous clean state */
	if (nospares
	    && (mddev->in_sync && mddev->recovery_cp == MaxSector)
	    && mddev->can_decrease_events
	    && mddev->events != 1) {
		mddev->events--;
		mddev->can_decrease_events = 0;
	} else {
		/* otherwise we have to go forward and ... */
		mddev->events ++;
		mddev->can_decrease_events = nospares;
	}

	/*
	 * This 64-bit counter should never wrap.
	 * Either we are in around ~1 trillion A.C., assuming
	 * 1 reboot per second, or we have a bug...
	 */
	WARN_ON(mddev->events == 0);

	rdev_for_each(rdev, mddev) {
		if (rdev->badblocks.changed)
			any_badblocks_changed++;
		if (test_bit(Faulty, &rdev->flags))
			set_bit(FaultRecorded, &rdev->flags);
	}

	sync_sbs(mddev, nospares);
	spin_unlock(&mddev->lock);

	pr_debug("md: updating %s RAID superblock on device (in sync %d)\n",
		 mdname(mddev), mddev->in_sync);

	mddev_add_trace_msg(mddev, "md md_update_sb");
rewrite:
	md_bitmap_update_sb(mddev->bitmap);
	rdev_for_each(rdev, mddev) {
		if (rdev->sb_loaded != 1)
			continue; /* no noise on spare devices */

		if (!test_bit(Faulty, &rdev->flags)) {
			md_super_write(mddev,rdev,
				       rdev->sb_start, rdev->sb_size,
				       rdev->sb_page);
			pr_debug("md: (write) %pg's sb offset: %llu\n",
				 rdev->bdev,
				 (unsigned long long)rdev->sb_start);
			rdev->sb_events = mddev->events;
			if (rdev->badblocks.size) {
				md_super_write(mddev, rdev,
					       rdev->badblocks.sector,
					       rdev->badblocks.size << 9,
					       rdev->bb_page);
				rdev->badblocks.size = 0;
			}

		} else
			pr_debug("md: %pg (skipping faulty)\n",
				 rdev->bdev);
	}
	if (md_super_wait(mddev) < 0)
		goto rewrite;
	/* if there was a failure, MD_SB_CHANGE_DEVS was set, and we re-write super */

	if (mddev_is_clustered(mddev) && ret == 0)
		md_cluster_ops->metadata_update_finish(mddev);

	if (mddev->in_sync != sync_req ||
	    !bit_clear_unless(&mddev->sb_flags, BIT(MD_SB_CHANGE_PENDING),
			       BIT(MD_SB_CHANGE_DEVS) | BIT(MD_SB_CHANGE_CLEAN)))
		/* have to write it out again */
		goto repeat;
	wake_up(&mddev->sb_wait);
	if (test_bit(MD_RECOVERY_RUNNING, &mddev->recovery))
		sysfs_notify_dirent_safe(mddev->sysfs_completed);

	rdev_for_each(rdev, mddev) {
		if (test_and_clear_bit(FaultRecorded, &rdev->flags))
			clear_bit(Blocked, &rdev->flags);

		if (any_badblocks_changed)
			ack_all_badblocks(&rdev->badblocks);
		clear_bit(BlockedBadBlocks, &rdev->flags);
		wake_up(&rdev->blocked_wait);
	}
}
EXPORT_SYMBOL(md_update_sb);

static int add_bound_rdev(struct md_rdev *rdev)
{
	struct mddev *mddev = rdev->mddev;
	int err = 0;
	bool add_journal = test_bit(Journal, &rdev->flags);

	if (!mddev->pers->hot_remove_disk || add_journal) {
		/* If there is hot_add_disk but no hot_remove_disk
		 * then added disks for geometry changes,
		 * and should be added immediately.
		 */
		super_types[mddev->major_version].
			validate_super(mddev, NULL/*freshest*/, rdev);
		err = mddev->pers->hot_add_disk(mddev, rdev);
		if (err) {
			md_kick_rdev_from_array(rdev);
			return err;
		}
	}
	sysfs_notify_dirent_safe(rdev->sysfs_state);

	set_bit(MD_SB_CHANGE_DEVS, &mddev->sb_flags);
	if (mddev->degraded)
		set_bit(MD_RECOVERY_RECOVER, &mddev->recovery);
	set_bit(MD_RECOVERY_NEEDED, &mddev->recovery);
	md_new_event();
	return 0;
}

/* words written to sysfs files may, or may not, be \n terminated.
 * We want to accept with case. For this we use cmd_match.
 */
static int cmd_match(const char *cmd, const char *str)
{
	/* See if cmd, written into a sysfs file, matches
	 * str.  They must either be the same, or cmd can
	 * have a trailing newline
	 */
	while (*cmd && *str && *cmd == *str) {
		cmd++;
		str++;
	}
	if (*cmd == '\n')
		cmd++;
	if (*str || *cmd)
		return 0;
	return 1;
}

struct rdev_sysfs_entry {
	struct attribute attr;
	ssize_t (*show)(struct md_rdev *, char *);
	ssize_t (*store)(struct md_rdev *, const char *, size_t);
};

static ssize_t
state_show(struct md_rdev *rdev, char *page)
{
	char *sep = ",";
	size_t len = 0;
	unsigned long flags = READ_ONCE(rdev->flags);

	if (test_bit(Faulty, &flags) ||
	    (!test_bit(ExternalBbl, &flags) &&