Merge tag 'md-6.15-20250312' of https://git.kernel.org/pub/scm/linux/kernel/git/mdraid/linux into for-6.15/block

Merge MD changes from Yu:

"- fix recovery can preempt resync (Li Nan)
 - fix md-bitmap IO limit (Su Yue)
 - fix raid10 discard with REQ_NOWAIT (Xiao Ni)
 - fix raid1 memory leak (Zheng Qixing)
 - fix mddev uaf (Yu Kuai)
 - fix raid1,raid10 IO flags (Yu Kuai)
 - some refactor and cleanup (Yu Kuai)"

* tag 'md-6.15-20250312' of https://git.kernel.org/pub/scm/linux/kernel/git/mdraid/linux:
  md/raid10: wait barrier before returning discard request with REQ_NOWAIT
  md/md-bitmap: fix wrong bitmap_limit for clustermd when write sb
  md/raid1,raid10: don't ignore IO flags
  md/raid5: merge reshape_progress checking inside get_reshape_loc()
  md: fix mddev uaf while iterating all_mddevs list
  md: switch md-cluster to use md_submodle_head
  md: don't export md_cluster_ops
  md/md-cluster: cleanup md_cluster_ops reference
  md: switch personalities to use md_submodule_head
  md: introduce struct md_submodule_head and APIs
  md: only include md-cluster.h if necessary
  md: merge common code into find_pers()
  md/raid1: fix memory leak in raid1_run() if no active rdev
  md: ensure resync is prioritized over recovery

11 files changed, 338 insertions, 269 deletions

diff --git a/drivers/md/md-bitmap.c b/drivers/md/md-bitmap.c
index 23c09d22fcdb..44ec9b17cfd3 100644
--- a/drivers/md/md-bitmap.c
+++ b/drivers/md/md-bitmap.c
@@ -29,8 +29,10 @@
 #include <linux/buffer_head.h>
 #include <linux/seq_file.h>
 #include <trace/events/block.h>
+
 #include "md.h"
 #include "md-bitmap.h"
+#include "md-cluster.h"
 
 #define BITMAP_MAJOR_LO 3
 /* version 4 insists the bitmap is in little-endian order
@@ -426,8 +428,8 @@ static int __write_sb_page(struct md_rdev *rdev, struct bitmap *bitmap,
 	struct block_device *bdev;
 	struct mddev *mddev = bitmap->mddev;
 	struct bitmap_storage *store = &bitmap->storage;
-	unsigned int bitmap_limit = (bitmap->storage.file_pages - pg_index) <<
-		PAGE_SHIFT;
+	unsigned long num_pages = bitmap->storage.file_pages;
+	unsigned int bitmap_limit = (num_pages - pg_index % num_pages) << PAGE_SHIFT;
 	loff_t sboff, offset = mddev->bitmap_info.offset;
 	sector_t ps = pg_index * PAGE_SIZE / SECTOR_SIZE;
 	unsigned int size = PAGE_SIZE;
@@ -436,7 +438,7 @@ static int __write_sb_page(struct md_rdev *rdev, struct bitmap *bitmap,
 
 	bdev = (rdev->meta_bdev) ? rdev->meta_bdev : rdev->bdev;
 	/* we compare length (page numbers), not page offset. */
-	if ((pg_index - store->sb_index) == store->file_pages - 1) {
+	if ((pg_index - store->sb_index) == num_pages - 1) {
 		unsigned int last_page_size = store->bytes & (PAGE_SIZE - 1);
 
 		if (last_page_size == 0)
@@ -942,7 +944,7 @@ out:
 				bmname(bitmap), err);
 			goto out_no_sb;
 		}
-		bitmap->cluster_slot = md_cluster_ops->slot_number(bitmap->mddev);
+		bitmap->cluster_slot = bitmap->mddev->cluster_ops->slot_number(bitmap->mddev);
 		goto re_read;
 	}
 
@@ -2021,7 +2023,7 @@ static void md_bitmap_free(void *data)
 		sysfs_put(bitmap->sysfs_can_clear);
 
 	if (mddev_is_clustered(bitmap->mddev) && bitmap->mddev->cluster_info &&
-		bitmap->cluster_slot == md_cluster_ops->slot_number(bitmap->mddev))
+		bitmap->cluster_slot == bitmap->mddev->cluster_ops->slot_number(bitmap->mddev))
 		md_cluster_stop(bitmap->mddev);
 
 	/* Shouldn't be needed - but just in case.... */
@@ -2229,7 +2231,7 @@ static int bitmap_load(struct mddev *mddev)
 		mddev_create_serial_pool(mddev, rdev);
 
 	if (mddev_is_clustered(mddev))
-		md_cluster_ops->load_bitmaps(mddev, mddev->bitmap_info.nodes);
+		mddev->cluster_ops->load_bitmaps(mddev, mddev->bitmap_info.nodes);
 
 	/* Clear out old bitmap info first:  Either there is none, or we
 	 * are resuming after someone else has possibly changed things,
diff --git a/drivers/md/md-cluster.c b/drivers/md/md-cluster.c
index 6595f89becdb..94221d964d4f 100644
--- a/drivers/md/md-cluster.c
+++ b/drivers/md/md-cluster.c
@@ -1166,7 +1166,7 @@ static int resize_bitmaps(struct mddev *mddev, sector_t newsize, sector_t oldsiz
 		struct dlm_lock_resource *bm_lockres;
 		char str[64];
 
-		if (i == md_cluster_ops->slot_number(mddev))
+		if (i == slot_number(mddev))
 			continue;
 
 		bitmap = mddev->bitmap_ops->get_from_slot(mddev, i);
@@ -1216,7 +1216,7 @@ out:
  */
 static int cluster_check_sync_size(struct mddev *mddev)
 {
-	int current_slot = md_cluster_ops->slot_number(mddev);
+	int current_slot = slot_number(mddev);
 	int node_num = mddev->bitmap_info.nodes;
 	struct dlm_lock_resource *bm_lockres;
 	struct md_bitmap_stats stats;
@@ -1612,7 +1612,14 @@ out:
 	return err;
 }
 
-static const struct md_cluster_operations cluster_ops = {
+static struct md_cluster_operations cluster_ops = {
+	.head = {
+		.type	= MD_CLUSTER,
+		.id	= ID_CLUSTER,
+		.name	= "cluster",
+		.owner	= THIS_MODULE,
+	},
+
 	.join   = join,
 	.leave  = leave,
 	.slot_number = slot_number,
@@ -1642,13 +1649,12 @@ static int __init cluster_init(void)
 {
 	pr_warn("md-cluster: support raid1 and raid10 (limited support)\n");
 	pr_info("Registering Cluster MD functions\n");
-	register_md_cluster_operations(&cluster_ops, THIS_MODULE);
-	return 0;
+	return register_md_submodule(&cluster_ops.head);
 }
 
 static void cluster_exit(void)
 {
-	unregister_md_cluster_operations();
+	unregister_md_submodule(&cluster_ops.head);
 }
 
 module_init(cluster_init);
diff --git a/drivers/md/md-cluster.h b/drivers/md/md-cluster.h
index 470bf18ffde5..8fb06d853173 100644
--- a/drivers/md/md-cluster.h
+++ b/drivers/md/md-cluster.h
@@ -10,6 +10,8 @@ struct mddev;
 struct md_rdev;
 
 struct md_cluster_operations {
+	struct md_submodule_head head;
+
 	int (*join)(struct mddev *mddev, int nodes);
 	int (*leave)(struct mddev *mddev);
 	int (*slot_number)(struct mddev *mddev);
@@ -35,4 +37,8 @@ struct md_cluster_operations {
 	void (*update_size)(struct mddev *mddev, sector_t old_dev_sectors);
 };
 
+extern int md_setup_cluster(struct mddev *mddev, int nodes);
+extern void md_cluster_stop(struct mddev *mddev);
+extern void md_reload_sb(struct mddev *mddev, int raid_disk);
+
 #endif /* _MD_CLUSTER_H */
diff --git a/drivers/md/md-linear.c b/drivers/md/md-linear.c
index 369aed044b40..5d9b08115375 100644
--- a/drivers/md/md-linear.c
+++ b/drivers/md/md-linear.c
@@ -5,7 +5,6 @@
  */
 
 #include <linux/blkdev.h>
-#include <linux/raid/md_u.h>
 #include <linux/seq_file.h>
 #include <linux/module.h>
 #include <linux/slab.h>
@@ -320,9 +319,13 @@ static void linear_quiesce(struct mddev *mddev, int state)
 }
 
 static struct md_personality linear_personality = {
-	.name		= "linear",
-	.level		= LEVEL_LINEAR,
-	.owner		= THIS_MODULE,
+	.head = {
+		.type	= MD_PERSONALITY,
+		.id	= ID_LINEAR,
+		.name	= "linear",
+		.owner	= THIS_MODULE,
+	},
+
 	.make_request	= linear_make_request,
 	.run		= linear_run,
 	.free		= linear_free,
@@ -335,12 +338,12 @@ static struct md_personality linear_personality = {
 
 static int __init linear_init(void)
 {
-	return register_md_personality(&linear_personality);
+	return register_md_submodule(&linear_personality.head);
 }
 
 static void linear_exit(void)
 {
-	unregister_md_personality(&linear_personality);
+	unregister_md_submodule(&linear_personality.head);
 }
 
 module_init(linear_init);
diff --git a/drivers/md/md.c b/drivers/md/md.c
index 4092fb94842d..438e71e45c16 100644
--- a/drivers/md/md.c
+++ b/drivers/md/md.c
@@ -79,16 +79,10 @@ static const char *action_name[NR_SYNC_ACTIONS] = {
 	[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 DEFINE_XARRAY(md_submodule);
 
 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;
 
@@ -629,6 +623,12 @@ static void __mddev_put(struct mddev *mddev)
 	queue_work(md_misc_wq, &mddev->del_work);
 }
 
+static void mddev_put_locked(struct mddev *mddev)
+{
+	if (atomic_dec_and_test(&mddev->active))
+		__mddev_put(mddev);
+}
+
 void mddev_put(struct mddev *mddev)
 {
 	if (!atomic_dec_and_lock(&mddev->active, &all_mddevs_lock))
@@ -888,16 +888,40 @@ struct md_rdev *md_find_rdev_rcu(struct mddev *mddev, dev_t dev)
 }
 EXPORT_SYMBOL_GPL(md_find_rdev_rcu);
 
-static struct md_personality *find_pers(int level, char *clevel)
+static struct md_personality *get_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;
+	struct md_personality *ret = NULL;
+	struct md_submodule_head *head;
+	unsigned long i;
+
+	xa_lock(&md_submodule);
+	xa_for_each(&md_submodule, i, head) {
+		if (head->type != MD_PERSONALITY)
+			continue;
+		if ((level != LEVEL_NONE && head->id == level) ||
+		    !strcmp(head->name, clevel)) {
+			if (try_module_get(head->owner))
+				ret = (void *)head;
+			break;
+		}
 	}
-	return NULL;
+	xa_unlock(&md_submodule);
+
+	if (!ret) {
+		if (level != LEVEL_NONE)
+			pr_warn("md: personality for level %d is not loaded!\n",
+				level);
+		else
+			pr_warn("md: personality for level %s is not loaded!\n",
+				clevel);
+	}
+
+	return ret;
+}
+
+static void put_pers(struct md_personality *pers)
+{
+	module_put(pers->head.owner);
 }
 
 /* return the offset of the super block in 512byte sectors */
@@ -1180,7 +1204,7 @@ 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);
+		mdname(mddev), mddev->pers->head.name);
 	return 1;
 }
 EXPORT_SYMBOL(md_check_no_bitmap);
@@ -2626,11 +2650,11 @@ repeat:
 			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);
+		ret = mddev->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);
+				mddev->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));
@@ -2770,7 +2794,7 @@ 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);
+		mddev->cluster_ops->metadata_update_finish(mddev);
 
 	if (mddev->in_sync != sync_req ||
 	    !bit_clear_unless(&mddev->sb_flags, BIT(MD_SB_CHANGE_PENDING),
@@ -2929,7 +2953,7 @@ state_store(struct md_rdev *rdev, const char *buf, size_t len)
 		else {
 			err = 0;
 			if (mddev_is_clustered(mddev))
-				err = md_cluster_ops->remove_disk(mddev, rdev);
+				err = mddev->cluster_ops->remove_disk(mddev, rdev);
 
 			if (err == 0) {
 				md_kick_rdev_from_array(rdev);
@@ -3039,7 +3063,7 @@ state_store(struct md_rdev *rdev, const char *buf, size_t len)
 			 * by this node eventually
 			 */
 			if (!mddev_is_clustered(rdev->mddev) ||
-			    (err = md_cluster_ops->gather_bitmaps(rdev)) == 0) {
+			    (err = mddev->cluster_ops->gather_bitmaps(rdev)) == 0) {
 				clear_bit(Faulty, &rdev->flags);
 				err = add_bound_rdev(rdev);
 			}
@@ -3847,7 +3871,7 @@ level_show(struct mddev *mddev, char *page)
 	spin_lock(&mddev->lock);
 	p = mddev->pers;
 	if (p)
-		ret = sprintf(page, "%s\n", p->name);
+		ret = sprintf(page, "%s\n", p->head.name);
 	else if (mddev->clevel[0])
 		ret = sprintf(page, "%s\n", mddev->clevel);
 	else if (mddev->level != LEVEL_NONE)
@@ -3904,7 +3928,7 @@ level_store(struct mddev *mddev, const char *buf, size_t len)
 	rv = -EINVAL;
 	if (!mddev->pers->quiesce) {
 		pr_warn("md: %s: %s does not support online personality change\n",
-			mdname(mddev), mddev->pers->name);
+			mdname(mddev), mddev->pers->head.name);
 		goto out_unlock;
 	}
 
@@ -3918,24 +3942,20 @@ level_store(struct mddev *mddev, const char *buf, size_t len)
 
 	if (request_module("md-%s", clevel) != 0)
 		request_module("md-level-%s", clevel);
-	spin_lock(&pers_lock);
-	pers = find_pers(level, clevel);
-	if (!pers || !try_module_get(pers->owner)) {
-		spin_unlock(&pers_lock);
-		pr_warn("md: personality %s not loaded\n", clevel);
+	pers = get_pers(level, clevel);
+	if (!pers) {
 		rv = -EINVAL;
 		goto out_unlock;
 	}
-	spin_unlock(&pers_lock);
 
 	if (pers == mddev->pers) {
 		/* Nothing to do! */
-		module_put(pers->owner);
+		put_pers(pers);
 		rv = len;
 		goto out_unlock;
 	}
 	if (!pers->takeover) {
-		module_put(pers->owner);
+		put_pers(pers);
 		pr_warn("md: %s: %s does not support personality takeover\n",
 			mdname(mddev), clevel);
 		rv = -EINVAL;
@@ -3956,7 +3976,7 @@ level_store(struct mddev *mddev, const char *buf, size_t len)
 		mddev->raid_disks -= mddev->delta_disks;
 		mddev->delta_disks = 0;
 		mddev->reshape_backwards = 0;
-		module_put(pers->owner);
+		put_pers(pers);
 		pr_warn("md: %s: %s would not accept array\n",
 			mdname(mddev), clevel);
 		rv = PTR_ERR(priv);
@@ -3971,7 +3991,7 @@ level_store(struct mddev *mddev, const char *buf, size_t len)
 	oldpriv = mddev->private;
 	mddev->pers = pers;
 	mddev->private = priv;
-	strscpy(mddev->clevel, pers->name, sizeof(mddev->clevel));
+	strscpy(mddev->clevel, pers->head.name, sizeof(mddev->clevel));
 	mddev->level = mddev->new_level;
 	mddev->layout = mddev->new_layout;
 	mddev->chunk_sectors = mddev->new_chunk_sectors;
@@ -4013,7 +4033,7 @@ level_store(struct mddev *mddev, const char *buf, size_t len)
 			mddev->to_remove = &md_redundancy_group;
 	}
 
-	module_put(oldpers->owner);
+	put_pers(oldpers);
 
 	rdev_for_each(rdev, mddev) {
 		if (rdev->raid_disk < 0)
@@ -5571,7 +5591,7 @@ __ATTR(fail_last_dev, S_IRUGO | S_IWUSR, fail_last_dev_show,
 
 static ssize_t serialize_policy_show(struct mddev *mddev, char *page)
 {
-	if (mddev->pers == NULL || (mddev->pers->level != 1))
+	if (mddev->pers == NULL || (mddev->pers->head.id != ID_RAID1))
 		return sprintf(page, "n/a\n");
 	else
 		return sprintf(page, "%d\n", mddev->serialize_policy);
@@ -5597,7 +5617,7 @@ serialize_policy_store(struct mddev *mddev, const char *buf, size_t len)
 	err = mddev_suspend_and_lock(mddev);
 	if (err)
 		return err;
-	if (mddev->pers == NULL || (mddev->pers->level != 1)) {
+	if (mddev->pers == NULL || (mddev->pers->head.id != ID_RAID1)) {
 		pr_err("md: serialize_policy is only effective for raid1\n");
 		err = -EINVAL;
 		goto unlock;
@@ -6083,30 +6103,21 @@ int md_run(struct mddev *mddev)
 			goto exit_sync_set;
 	}
 
-	spin_lock(&pers_lock);
-	pers = find_pers(mddev->level, mddev->clevel);
-	if (!pers || !try_module_get(pers->owner)) {
-		spin_unlock(&pers_lock);
-		if (mddev->level != LEVEL_NONE)
-			pr_warn("md: personality for level %d is not loaded!\n",
-				mddev->level);
-		else
-			pr_warn("md: personality for level %s is not loaded!\n",
-				mddev->clevel);
+	pers = get_pers(mddev->level, mddev->clevel);
+	if (!pers) {
 		err = -EINVAL;
 		goto abort;
 	}
-	spin_unlock(&pers_lock);
-	if (mddev->level != pers->level) {
-		mddev->level = pers->level;
-		mddev->new_level = pers->level;
+	if (mddev->level != pers->head.id) {
+		mddev->level = pers->head.id;
+		mddev->new_level = pers->head.id;
 	}
-	strscpy(mddev->clevel, pers->name, sizeof(mddev->clevel));
+	strscpy(mddev->clevel, pers->head.name, sizeof(mddev->clevel));
 
 	if (mddev->reshape_position != MaxSector &&
 	    pers->start_reshape == NULL) {
 		/* This personality cannot handle reshaping... */
-		module_put(pers->owner);
+		put_pers(pers);
 		err = -EINVAL;
 		goto abort;
 	}
@@ -6233,7 +6244,7 @@ bitmap_abort:
 	if (mddev->private)
 		pers->free(mddev, mddev->private);
 	mddev->private = NULL;
-	module_put(pers->owner);
+	put_pers(pers);
 	mddev->bitmap_ops->destroy(mddev);
 abort:
 	bioset_exit(&mddev->io_clone_set);
@@ -6454,7 +6465,7 @@ static void __md_stop(struct mddev *mddev)
 	mddev->private = NULL;
 	if (pers->sync_request && mddev->to_remove == NULL)
 		mddev->to_remove = &md_redundancy_group;
-	module_put(pers->owner);
+	put_pers(pers);
 	clear_bit(MD_RECOVERY_FROZEN, &mddev->recovery);
 
 	bioset_exit(&mddev->bio_set);
@@ -6970,7 +6981,7 @@ int md_add_new_disk(struct mddev *mddev, struct mdu_disk_info_s *info)
 				set_bit(Candidate, &rdev->flags);
 			else if (info->state & (1 << MD_DISK_CLUSTER_ADD)) {
 				/* --add initiated by this node */
-				err = md_cluster_ops->add_new_disk(mddev, rdev);
+				err = mddev->cluster_ops->add_new_disk(mddev, rdev);
 				if (err) {
 					export_rdev(rdev, mddev);
 					return err;
@@ -6987,14 +6998,14 @@ int md_add_new_disk(struct mddev *mddev, struct mdu_disk_info_s *info)
 		if (mddev_is_clustered(mddev)) {
 			if (info->state & (1 << MD_DISK_CANDIDATE)) {
 				if (!err) {
-					err = md_cluster_ops->new_disk_ack(mddev,
-						err == 0);
+					err = mddev->cluster_ops->new_disk_ack(
+							mddev, err == 0);
 					if (err)
 						md_kick_rdev_from_array(rdev);
 				}
 			} else {
 				if (err)
-					md_cluster_ops->add_new_disk_cancel(mddev);
+					mddev->cluster_ops->add_new_disk_cancel(mddev);
 				else
 					err = add_bound_rdev(rdev);
 			}
@@ -7074,10 +7085,9 @@ static int hot_remove_disk(struct mddev *mddev, dev_t dev)
 		goto busy;
 
 kick_rdev:
-	if (mddev_is_clustered(mddev)) {
-		if (md_cluster_ops->remove_disk(mddev, rdev))
-			goto busy;
-	}
+	if (mddev_is_clustered(mddev) &&
+	    mddev->cluster_ops->remove_disk(mddev, rdev))
+		goto busy;
 
 	md_kick_rdev_from_array(rdev);
 	set_bit(MD_SB_CHANGE_DEVS, &mddev->sb_flags);
@@ -7380,7 +7390,7 @@ static int update_size(struct mddev *mddev, sector_t num_sectors)
 	rv = mddev->pers->resize(mddev, num_sectors);
 	if (!rv) {
 		if (mddev_is_clustered(mddev))
-			md_cluster_ops->update_size(mddev, old_dev_sectors);
+			mddev->cluster_ops->update_size(mddev, old_dev_sectors);
 		else if (!mddev_is_dm(mddev))
 			set_capacity_and_notify(mddev->gendisk,
 						mddev->array_sectors);
@@ -7428,6 +7438,28 @@ static int update_raid_disks(struct mddev *mddev, int raid_disks)
 	return rv;
 }
 
+static int get_cluster_ops(struct mddev *mddev)
+{
+	xa_lock(&md_submodule);
+	mddev->cluster_ops = xa_load(&md_submodule, ID_CLUSTER);
+	if (mddev->cluster_ops &&
+	    !try_module_get(mddev->cluster_ops->head.owner))
+		mddev->cluster_ops = NULL;
+	xa_unlock(&md_submodule);
+
+	return mddev->cluster_ops == NULL ? -ENOENT : 0;
+}
+
+static void put_cluster_ops(struct mddev *mddev)
+{
+	if (!mddev->cluster_ops)
+		return;
+
+	mddev->cluster_ops->leave(mddev);
+	module_put(mddev->cluster_ops->head.owner);
+	mddev->cluster_ops = NULL;
+}
+
 /*
  * update_array_info is used to change the configuration of an
  * on-line array.
@@ -7536,16 +7568,15 @@ static int update_array_info(struct mddev *mddev, mdu_array_info_t *info)
 
 			if (mddev->bitmap_info.nodes) {
 				/* hold PW on all the bitmap lock */
-				if (md_cluster_ops->lock_all_bitmaps(mddev) <= 0) {
+				if (mddev->cluster_ops->lock_all_bitmaps(mddev) <= 0) {
 					pr_warn("md: can't change bitmap to none since the array is in use by more than one node\n");
 					rv = -EPERM;
-					md_cluster_ops->unlock_all_bitmaps(mddev);
+					mddev->cluster_ops->unlock_all_bitmaps(mddev);
 					goto err;
 				}
 
 				mddev->bitmap_info.nodes = 0;
-				md_cluster_ops->leave(mddev);
-				module_put(md_cluster_mod);
+				put_cluster_ops(mddev);
 				mddev->safemode_delay = DEFAULT_SAFEMODE_DELAY;
 			}
 			mddev->bitmap_ops->destroy(mddev);
@@ -7829,7 +7860,7 @@ static int md_ioctl(struct block_device *bdev, blk_mode_t mode,
 
 	case CLUSTERED_DISK_NACK:
 		if (mddev_is_clustered(mddev))
-			md_cluster_ops->new_disk_ack(mddev, false);
+			mddev->cluster_ops->new_disk_ack(mddev, false);
 		else
 			err = -EINVAL;
 		goto unlock;
@@ -8111,7 +8142,8 @@ void md_error(struct mddev *mddev, struct md_rdev *rdev)
 		return;
 	mddev->pers->error_handler(mddev, rdev);
 
-	if (mddev->pers->level == 0 || mddev->pers->level == LEVEL_LINEAR)
+	if (mddev->pers->head.id == ID_RAID0 ||
+	    mddev->pers->head.id == ID_LINEAR)
 		return;
 
 	if (mddev->degraded && !test_bit(MD_BROKEN, &mddev->flags))
@@ -8149,14 +8181,17 @@ static void status_unused(struct seq_file *seq)
 
 static void status_personalities(struct seq_file *seq)
 {
-	struct md_personality *pers;
+	struct md_submodule_head *head;
+	unsigned long i;
 
 	seq_puts(seq, "Personalities : ");
-	spin_lock(&pers_lock);
-	list_for_each_entry(pers, &pers_list, list)
-		seq_printf(seq, "[%s] ", pers->name);
 
-	spin_unlock(&pers_lock);
+	xa_lock(&md_submodule);
+	xa_for_each(&md_submodule, i, head)
+		if (head->type == MD_PERSONALITY)
+			seq_printf(seq, "[%s] ", head->name);
+	xa_unlock(&md_submodule);
+
 	seq_puts(seq, "\n");
 }
 
@@ -8379,7 +8414,7 @@ static int md_seq_show(struct seq_file *seq, void *v)
 				seq_printf(seq, " (read-only)");
 			if (mddev->ro == MD_AUTO_READ)
 				seq_printf(seq, " (auto-read-only)");
-			seq_printf(seq, " %s", mddev->pers->name);
+			seq_printf(seq, " %s", mddev->pers->head.name);
 		} else {
 			seq_printf(seq, "inactive");
 		}
@@ -8448,9 +8483,7 @@ static int md_seq_show(struct seq_file *seq, void *v)
 	if (mddev == list_last_entry(&all_mddevs, struct mddev, all_mddevs))
 		status_unused(seq);
 
-	if (atomic_dec_and_test(&mddev->active))
-		__mddev_put(mddev);
-
+	mddev_put_locked(mddev);
 	return 0;
 }
 
@@ -8501,67 +8534,34 @@ static const struct proc_ops mdstat_proc_ops = {
 	.proc_poll	= mdstat_poll,
 };
 
-int register_md_personality(struct md_personality *p)
-{
-	pr_debug("md: %s personality registered for level %d\n",
-		 p->name, p->level);
-	spin_lock(&pers_lock);
-	list_add_tail(&p->list, &pers_list);
-	spin_unlock(&pers_lock);
-	return 0;
-}
-EXPORT_SYMBOL(register_md_personality);
-
-int unregister_md_personality(struct md_personality *p)
-{
-	pr_debug("md: %s personality unregistered\n", p->name);
-	spin_lock(&pers_lock);
-	list_del_init(&p->list);
-	spin_unlock(&pers_lock);
-	return 0;
-}
-EXPORT_SYMBOL(unregister_md_personality);
-
-int register_md_cluster_operations(const struct md_cluster_operations *ops,
-				   struct module *module)
+int register_md_submodule(struct md_submodule_head *msh)
 {
-	int ret = 0;
-	spin_lock(&pers_lock);
-	if (md_cluster_ops != NULL)
-		ret = -EALREADY;
-	else {
-		md_cluster_ops = ops;
-		md_cluster_mod = module;
-	}
-	spin_unlock(&pers_lock);
-	return ret;
+	return xa_insert(&md_submodule, msh->id, msh, GFP_KERNEL);
 }
-EXPORT_SYMBOL(register_md_cluster_operations);
+EXPORT_SYMBOL_GPL(register_md_submodule);
 
-int unregister_md_cluster_operations(void)
+void unregister_md_submodule(struct md_submodule_head *msh)
 {
-	spin_lock(&pers_lock);
-	md_cluster_ops = NULL;
-	spin_unlock(&pers_lock);
-	return 0;
+	xa_erase(&md_submodule, msh->id);
 }
-EXPORT_SYMBOL(unregister_md_cluster_operations);
+EXPORT_SYMBOL_GPL(unregister_md_submodule);
 
 int md_setup_cluster(struct mddev *mddev, int nodes)
 {
-	int ret;
-	if (!md_cluster_ops)
+	int ret = get_cluster_ops(mddev);
+
+	if (ret) {
 		request_module("md-cluster");
-	spin_lock(&pers_lock);
+		ret = get_cluster_ops(mddev);
+	}
+
 	/* ensure module won't be unloaded */
-	if (!md_cluster_ops || !try_module_get(md_cluster_mod)) {
+	if (ret) {
 		pr_warn("can't find md-cluster module or get its reference.\n");
-		spin_unlock(&pers_lock);
-		return -ENOENT;
+		return ret;
 	}
-	spin_unlock(&pers_lock);
 
-	ret = md_cluster_ops->join(mddev, nodes);
+	ret = mddev->cluster_ops->join(mddev, nodes);
 	if (!ret)
 		mddev->safemode_delay = 0;
 	return ret;
@@ -8569,10 +8569,7 @@ int md_setup_cluster(struct mddev *mddev, int nodes)
 
 void md_cluster_stop(struct mddev *mddev)
 {
-	if (!md_cluster_ops)
-		return;
-	md_cluster_ops->leave(mddev);
-	module_put(md_cluster_mod);
+	put_cluster_ops(mddev);
 }
 
 static int is_mddev_idle(struct mddev *mddev, int init)
@@ -8965,7 +8962,7 @@ void md_do_sync(struct md_thread *thread)
 	}
 
 	if (mddev_is_clustered(mddev)) {
-		ret = md_cluster_ops->resync_start(mddev);
+		ret = mddev->cluster_ops->resync_start(mddev);
 		if (ret)
 			goto skip;
 
@@ -8992,7 +8989,7 @@ void md_do_sync(struct md_thread *thread)
 	 *
 	 */
 	if (mddev_is_clustered(mddev))
-		md_cluster_ops->resync_start_notify(mddev);
+		mddev->cluster_ops->resync_start_notify(mddev);
 	do {
 		int mddev2_minor = -1;
 		mddev->curr_resync = MD_RESYNC_DELAYED;
@@ -9447,6 +9444,13 @@ static bool md_choose_sync_action(struct mddev *mddev, int *spares)
 		return true;
 	}
 
+	/* Check if resync is in progress. */
+	if (mddev->recovery_cp < MaxSector) {
+		set_bit(MD_RECOVERY_SYNC, &mddev->recovery);
+		clear_bit(MD_RECOVERY_RECOVER, &mddev->recovery);
+		return true;
+	}
+
 	/*
 	 * Remove any failed drives, then add spares if possible. Spares are
 	 * also removed and re-added, to allow the personality to fail the
@@ -9463,13 +9467,6 @@ static bool md_choose_sync_action(struct mddev *mddev, int *spares)
 		return true;
 	}
 
-	/* Check if recovery is in progress. */
-	if (mddev->recovery_cp < MaxSector) {
-		set_bit(MD_RECOVERY_SYNC, &mddev->recovery);
-		clear_bit(MD_RECOVERY_RECOVER, &mddev->recovery);
-		return true;
-	}
-
 	/* Delay to choose resync/check/repair in md_do_sync(). */
 	if (test_bit(MD_RECOVERY_SYNC, &mddev->recovery))
 		return true;
@@ -9776,7 +9773,7 @@ void md_reap_sync_thread(struct mddev *mddev)
 	 * call resync_finish here if MD_CLUSTER_RESYNC_LOCKED is set by
 	 * clustered raid */
 	if (test_and_clear_bit(MD_CLUSTER_RESYNC_LOCKED, &mddev->flags))
-		md_cluster_ops->resync_finish(mddev);
+		mddev->cluster_ops->resync_finish(mddev);
 	clear_bit(MD_RECOVERY_RUNNING, &mddev->recovery);
 	clear_bit(MD_RECOVERY_DONE, &mddev->recovery);
 	clear_bit(MD_RECOVERY_SYNC, &mddev->recovery);
@@ -9784,13 +9781,13 @@ void md_reap_sync_thread(struct mddev *mddev)
 	clear_bit(MD_RECOVERY_REQUESTED, &mddev->recovery);
 	clear_bit(MD_RECOVERY_CHECK, &mddev->recovery);
 	/*
-	 * We call md_cluster_ops->update_size here because sync_size could
+	 * We call mddev->cluster_ops->update_size here because sync_size could
 	 * be changed by md_update_sb, and MD_RECOVERY_RESHAPE is cleared,
 	 * so it is time to update size across cluster.
 	 */
 	if (mddev_is_clustered(mddev) && is_reshaped
 				      && !test_bit(MD_CLOSING, &mddev->flags))
-		md_cluster_ops->update_size(mddev, old_dev_sectors);
+		mddev->cluster_ops->update_size(mddev, old_dev_sectors);
 	/* flag recovery needed just to double check */
 	set_bit(MD_RECOVERY_NEEDED, &mddev->recovery);
 	sysfs_notify_dirent_safe(mddev->sysfs_completed);
@@ -9882,11 +9879,11 @@ EXPORT_SYMBOL_GPL(rdev_clear_badblocks);
 static int md_notify_reboot(struct notifier_block *this,
 			    unsigned long code, void *x)
 {
-	struct mddev *mddev, *n;
+	struct mddev *mddev;
 	int need_delay = 0;
 
 	spin_lock(&all_mddevs_lock);
-	list_for_each_entry_safe(mddev, n, &all_mddevs, all_mddevs) {
+	list_for_each_entry(mddev, &all_mddevs, all_mddevs) {
 		if (!mddev_get(mddev))
 			continue;
 		spin_unlock(&all_mddevs_lock);
@@ -9898,8 +9895,8 @@ static int md_notify_reboot(struct notifier_block *this,
 			mddev_unlock(mddev);
 		}
 		need_delay = 1;
-		mddev_put(mddev);
 		spin_lock(&all_mddevs_lock);
+		mddev_put_locked(mddev);
 	}
 	spin_unlock(&all_mddevs_lock);
 
@@ -10016,7 +10013,7 @@ static void check_sb_changes(struct mddev *mddev, struct md_rdev *rdev)
 			if (rdev2->raid_disk == -1 && role != MD_DISK_ROLE_SPARE &&
 			    !(le32_to_cpu(sb->feature_map) &
 			      MD_FEATURE_RESHAPE_ACTIVE) &&
-			    !md_cluster_ops->resync_status_get(mddev)) {
+			    !mddev->cluster_ops->resync_status_get(mddev)) {
 				/*
 				 * -1 to make raid1_add_disk() set conf->fullsync
 				 * to 1. This could avoid skipping sync when the
@@ -10232,7 +10229,7 @@ void md_autostart_arrays(int part)
 
 static __exit void md_exit(void)
 {
-	struct mddev *mddev, *n;
+	struct mddev *mddev;
 	int delay = 1;
 
 	unregister_blkdev(MD_MAJOR,"md");
@@ -10253,7 +10250,7 @@ static __exit void md_exit(void)
 	remove_proc_entry("mdstat", NULL);
 
 	spin_lock(&all_mddevs_lock);
-	list_for_each_entry_safe(mddev, n, &all_mddevs, all_mddevs) {
+	list_for_each_entry(mddev, &all_mddevs, all_mddevs) {
 		if (!mddev_get(mddev))
 			continue;
 		spin_unlock(&all_mddevs_lock);
@@ -10265,8 +10262,8 @@ static __exit void md_exit(void)
 		 * the mddev for destruction by a workqueue, and the
 		 * destroy_workqueue() below will wait for that to complete.
 		 */
-		mddev_put(mddev);
 		spin_lock(&all_mddevs_lock);
+		mddev_put_locked(mddev);
 	}
 	spin_unlock(&all_mddevs_lock);
 
diff --git a/drivers/md/md.h b/drivers/md/md.h
index 6edc0f71b7d4..1cf00a04bcdd 100644
--- a/drivers/md/md.h
+++ b/drivers/md/md.h
@@ -18,11 +18,37 @@
 #include <linux/timer.h>
 #include <linux/wait.h>
 #include <linux/workqueue.h>
+#include <linux/raid/md_u.h>
 #include <trace/events/block.h>
-#include "md-cluster.h"
 
 #define MaxSector (~(sector_t)0)
 
+enum md_submodule_type {
+	MD_PERSONALITY = 0,
+	MD_CLUSTER,
+	MD_BITMAP, /* TODO */
+};
+
+enum md_submodule_id {
+	ID_LINEAR	= LEVEL_LINEAR,
+	ID_RAID0	= 0,
+	ID_RAID1	= 1,
+	ID_RAID4	= 4,
+	ID_RAID5	= 5,
+	ID_RAID6	= 6,
+	ID_RAID10	= 10,
+	ID_CLUSTER,
+	ID_BITMAP,	/* TODO */
+	ID_LLBITMAP,	/* TODO */
+};
+
+struct md_submodule_head {
+	enum md_submodule_type	type;
+	enum md_submodule_id	id;
+	const char		*name;
+	struct module		*owner;
+};
+
 /*
  * These flags should really be called "NO_RETRY" rather than
  * "FAILFAST" because they don't make any promise about time lapse,
@@ -294,6 +320,7 @@ extern bool rdev_set_badblocks(struct md_rdev *rdev, sector_t s, int sectors,
 extern void rdev_clear_badblocks(struct md_rdev *rdev, sector_t s, int sectors,
 				 int is_new);
 struct md_cluster_info;
+struct md_cluster_operations;
 
 /**
  * enum mddev_flags - md device flags.
@@ -576,6 +603,7 @@ struct mddev {
 	mempool_t *serial_info_pool;
 	void (*sync_super)(struct mddev *mddev, struct md_rdev *rdev);
 	struct md_cluster_info		*cluster_info;
+	struct md_cluster_operations *cluster_ops;
 	unsigned int			good_device_nr;	/* good device num within cluster raid */
 	unsigned int			noio_flag; /* for memalloc scope API */
 
@@ -699,10 +727,8 @@ static inline void md_sync_acct_bio(struct bio *bio, unsigned long nr_sectors)
 
 struct md_personality
 {
-	char *name;
-	int level;
-	struct list_head list;
-	struct module *owner;
+	struct md_submodule_head head;
+
 	bool __must_check (*make_request)(struct mddev *mddev, struct bio *bio);
 	/*
 	 * start up works that do NOT require md_thread. tasks that
@@ -843,13 +869,9 @@ static inline void safe_put_page(struct page *p)
 	if (p) put_page(p);
 }
 
-extern int register_md_personality(struct md_personality *p);
-extern int unregister_md_personality(struct md_personality *p);
-extern int register_md_cluster_operations(const struct md_cluster_operations *ops,
-		struct module *module);
-extern int unregister_md_cluster_operations(void);
-extern int md_setup_cluster(struct mddev *mddev, int nodes);
-extern void md_cluster_stop(struct mddev *mddev);
+int register_md_submodule(struct md_submodule_head *msh);
+void unregister_md_submodule(struct md_submodule_head *msh);
+
 extern struct md_thread *md_register_thread(
 	void (*run)(struct md_thread *thread),
 	struct mddev *mddev,
@@ -906,7 +928,6 @@ extern void md_idle_sync_thread(struct mddev *mddev);
 extern void md_frozen_sync_thread(struct mddev *mddev);
 extern void md_unfrozen_sync_thread(struct mddev *mddev);
 
-extern void md_reload_sb(struct mddev *mddev, int raid_disk);
 extern void md_update_sb(struct mddev *mddev, int force);
 extern void mddev_create_serial_pool(struct mddev *mddev, struct md_rdev *rdev);
 extern void mddev_destroy_serial_pool(struct mddev *mddev,
@@ -928,7 +949,6 @@ static inline void rdev_dec_pending(struct md_rdev *rdev, struct mddev *mddev)
 	}
 }
 
-extern const struct md_cluster_operations *md_cluster_ops;
 static inline int mddev_is_clustered(struct mddev *mddev)
 {
 	return mddev->cluster_info && mddev->bitmap_info.nodes > 1;
diff --git a/drivers/md/raid0.c b/drivers/md/raid0.c
index 8fc9339b00c7..2aec92e6e0a9 100644
--- a/drivers/md/raid0.c
+++ b/drivers/md/raid0.c
@@ -811,9 +811,13 @@ static void raid0_quiesce(struct mddev *mddev, int quiesce)
 
 static struct md_personality raid0_personality=
 {
-	.name		= "raid0",
-	.level		= 0,
-	.owner		= THIS_MODULE,
+	.head = {
+		.type	= MD_PERSONALITY,
+		.id	= ID_RAID0,
+		.name	= "raid0",
+		.owner	= THIS_MODULE,
+	},
+
 	.make_request	= raid0_make_request,
 	.run		= raid0_run,
 	.free		= raid0_free,
@@ -824,14 +828,14 @@ static struct md_personality raid0_personality=
 	.error_handler	= raid0_error,
 };
 
-static int __init raid0_init (void)
+static int __init raid0_init(void)
 {
-	return register_md_personality (&raid0_personality);
+	return register_md_submodule(&raid0_personality.head);
 }
 
-static void raid0_exit (void)
+static void __exit raid0_exit(void)
 {
-	unregister_md_personality (&raid0_personality);
+	unregister_md_submodule(&raid0_personality.head);
 }
 
 module_init(raid0_init);
diff --git a/drivers/md/raid1-10.c b/drivers/md/raid1-10.c
index 62b980b12f93..c7efd8aab675 100644
--- a/drivers/md/raid1-10.c
+++ b/drivers/md/raid1-10.c
@@ -287,8 +287,8 @@ static inline bool raid1_should_read_first(struct mddev *mddev,
 		return true;
 
 	if (mddev_is_clustered(mddev) &&
-	    md_cluster_ops->area_resyncing(mddev, READ, this_sector,
-					   this_sector + len))
+	    mddev->cluster_ops->area_resyncing(mddev, READ, this_sector,
+					       this_sector + len))
 		return true;
 
 	return false;
diff --git a/drivers/md/raid1.c b/drivers/md/raid1.c
index 0b2839105857..f06effec0a15 100644
--- a/drivers/md/raid1.c
+++ b/drivers/md/raid1.c
@@ -36,6 +36,7 @@
 #include "md.h"
 #include "raid1.h"
 #include "md-bitmap.h"
+#include "md-cluster.h"
 
 #define UNSUPPORTED_MDDEV_FLAGS		\
 	((1L << MD_HAS_JOURNAL) |	\
@@ -45,6 +46,7 @@
 
 static void allow_barrier(struct r1conf *conf, sector_t sector_nr);
 static void lower_barrier(struct r1conf *conf, sector_t sector_nr);
+static void raid1_free(struct mddev *mddev, void *priv);
 
 #define RAID_1_10_NAME "raid1"
 #include "raid1-10.c"
@@ -1315,8 +1317,6 @@ static void raid1_read_request(struct mddev *mddev, struct bio *bio,
 	struct r1conf *conf = mddev->private;
 	struct raid1_info *mirror;
 	struct bio *read_bio;
-	const enum req_op op = bio_op(bio);
-	const blk_opf_t do_sync = bio->bi_opf & REQ_SYNC;
 	int max_sectors;
 	int rdisk, error;
 	bool r1bio_existed = !!r1_bio;
@@ -1404,7 +1404,6 @@ static void raid1_read_request(struct mddev *mddev, struct bio *bio,
 	read_bio->bi_iter.bi_sector = r1_bio->sector +
 		mirror->rdev->data_offset;
 	read_bio->bi_end_io = raid1_end_read_request;
-	read_bio->bi_opf = op | do_sync;
 	if (test_bit(FailFast, &mirror->rdev->flags) &&
 	    test_bit(R1BIO_FailFast, &r1_bio->state))
 	        read_bio->bi_opf |= MD_FAILFAST;
@@ -1467,7 +1466,7 @@ static void raid1_write_request(struct mddev *mddev, struct bio *bio,
 	bool is_discard = (bio_op(bio) == REQ_OP_DISCARD);
 
 	if (mddev_is_clustered(mddev) &&
-	     md_cluster_ops->area_resyncing(mddev, WRITE,
+	    mddev->cluster_ops->area_resyncing(mddev, WRITE,
 		     bio->bi_iter.bi_sector, bio_end_sector(bio))) {
 
 		DEFINE_WAIT(w);
@@ -1478,7 +1477,7 @@ static void raid1_write_request(struct mddev *mddev, struct bio *bio,
 		for (;;) {
 			prepare_to_wait(&conf->wait_barrier,
 					&w, TASK_IDLE);
-			if (!md_cluster_ops->area_resyncing(mddev, WRITE,
+			if (!mddev->cluster_ops->area_resyncing(mddev, WRITE,
 							bio->bi_iter.bi_sector,
 							bio_end_sector(bio)))
 				break;
@@ -1653,8 +1652,6 @@ static void raid1_write_request(struct mddev *mddev, struct bio *bio,
 
 		mbio->bi_iter.bi_sector	= (r1_bio->sector + rdev->data_offset);
 		mbio->bi_end_io	= raid1_end_write_request;
-		mbio->bi_opf = bio_op(bio) |
-			(bio->bi_opf & (REQ_SYNC | REQ_FUA | REQ_ATOMIC));
 		if (test_bit(FailFast, &rdev->flags) &&
 		    !test_bit(WriteMostly, &rdev->flags) &&
 		    conf->raid_disks - mddev->degraded > 1)
@@ -3038,9 +3035,9 @@ static sector_t raid1_sync_request(struct mddev *mddev, sector_t sector_nr,
 		conf->cluster_sync_low = mddev->curr_resync_completed;
 		conf->cluster_sync_high = conf->cluster_sync_low + CLUSTER_RESYNC_WINDOW_SECTORS;
 		/* Send resync message */
-		md_cluster_ops->resync_info_update(mddev,
-				conf->cluster_sync_low,
-				conf->cluster_sync_high);
+		mddev->cluster_ops->resync_info_update(mddev,
+						       conf->cluster_sync_low,
+						       conf->cluster_sync_high);
 	}
 
 	/* For a user-requested sync, we read all readable devices and do a
@@ -3258,8 +3255,11 @@ static int raid1_run(struct mddev *mddev)
 
 	if (!mddev_is_dm(mddev)) {
 		ret = raid1_set_limits(mddev);
-		if (ret)
+		if (ret) {
+			if (!mddev->private)
+				raid1_free(mddev, conf);
 			return ret;
+		}
 	}
 
 	mddev->degraded = 0;
@@ -3273,6 +3273,8 @@ static int raid1_run(struct mddev *mddev)
 	 */
 	if (conf->raid_disks - mddev->degraded < 1) {
 		md_unregister_thread(mddev, &conf->thread);
+		if (!mddev->private)
+			raid1_free(mddev, conf);
 		return -EINVAL;
 	}
 
@@ -3493,9 +3495,13 @@ static void *raid1_takeover(struct mddev *mddev)
 
 static struct md_personality raid1_personality =
 {
-	.name		= "raid1",
-	.level		= 1,
-	.owner		= THIS_MODULE,
+	.head = {
+		.type	= MD_PERSONALITY,
+		.id	= ID_RAID1,
+		.name	= "raid1",
+		.owner	= THIS_MODULE,
+	},
+
 	.make_request	= raid1_make_request,
 	.run		= raid1_run,
 	.free		= raid1_free,
@@ -3512,18 +3518,18 @@ static struct md_personality raid1_personality =
 	.takeover	= raid1_takeover,
 };
 
-static int __init raid_init(void)
+static int __init raid1_init(void)
 {
-	return register_md_personality(&raid1_personality);
+	return register_md_submodule(&raid1_personality.head);
 }
 
-static void raid_exit(void)
+static void __exit raid1_exit(void)
 {
-	unregister_md_personality(&raid1_personality);
+	unregister_md_submodule(&raid1_personality.head);
 }
 
-module_init(raid_init);
-module_exit(raid_exit);
+module_init(raid1_init);
+module_exit(raid1_exit);
 MODULE_LICENSE("GPL");
 MODULE_DESCRIPTION("RAID1 (mirroring) personality for MD");
 MODULE_ALIAS("md-personality-3"); /* RAID1 */
diff --git a/drivers/md/raid10.c b/drivers/md/raid10.c
index a8664e29aada..d412aeeafc18 100644
--- a/drivers/md/raid10.c
+++ b/drivers/md/raid10.c
@@ -24,6 +24,7 @@
 #include "raid10.h"
 #include "raid0.h"
 #include "md-bitmap.h"
+#include "md-cluster.h"
 
 /*
  * RAID10 provides a combination of RAID0 and RAID1 functionality.
@@ -1146,8 +1147,6 @@ static void raid10_read_request(struct mddev *mddev, struct bio *bio,
 {
 	struct r10conf *conf = mddev->private;
 	struct bio *read_bio;
-	const enum req_op op = bio_op(bio);
-	const blk_opf_t do_sync = bio->bi_opf & REQ_SYNC;
 	int max_sectors;
 	struct md_rdev *rdev;
 	char b[BDEVNAME_SIZE];
@@ -1228,7 +1227,6 @@ static void raid10_read_request(struct mddev *mddev, struct bio *bio,
 	read_bio->bi_iter.bi_sector = r10_bio->devs[slot].addr +
 		choose_data_offset(r10_bio, rdev);
 	read_bio->bi_end_io = raid10_end_read_request;
-	read_bio->bi_opf = op | do_sync;
 	if (test_bit(FailFast, &rdev->flags) &&
 	    test_bit(R10BIO_FailFast, &r10_bio->state))
 	        read_bio->bi_opf |= MD_FAILFAST;
@@ -1247,10 +1245,6 @@ static void raid10_write_one_disk(struct mddev *mddev, struct r10bio *r10_bio,
 				  struct bio *bio, bool replacement,
 				  int n_copy)
 {
-	const enum req_op op = bio_op(bio);
-	const blk_opf_t do_sync = bio->bi_opf & REQ_SYNC;
-	const blk_opf_t do_fua = bio->bi_opf & REQ_FUA;
-	const blk_opf_t do_atomic = bio->bi_opf & REQ_ATOMIC;
 	unsigned long flags;
 	struct r10conf *conf = mddev->private;
 	struct md_rdev *rdev;
@@ -1269,7 +1263,6 @@ static void raid10_write_one_disk(struct mddev *mddev, struct r10bio *r10_bio,
 	mbio->bi_iter.bi_sector	= (r10_bio->devs[n_copy].addr +
 				   choose_data_offset(r10_bio, rdev));
 	mbio->bi_end_io	= raid10_end_write_request;
-	mbio->bi_opf = op | do_sync | do_fua | do_atomic;
 	if (!replacement && test_bit(FailFast,
 				     &conf->mirrors[devnum].rdev->flags)
 			 && enough(conf, devnum))
@@ -1355,9 +1348,9 @@ static void raid10_write_request(struct mddev *mddev, struct bio *bio,
 	int error;
 
 	if ((mddev_is_clustered(mddev) &&
-	     md_cluster_ops->area_resyncing(mddev, WRITE,
-					    bio->bi_iter.bi_sector,
-					    bio_end_sector(bio)))) {
+	     mddev->cluster_ops->area_resyncing(mddev, WRITE,
+						bio->bi_iter.bi_sector,
+						bio_end_sector(bio)))) {
 		DEFINE_WAIT(w);
 		/* Bail out if REQ_NOWAIT is set for the bio */
 		if (bio->bi_opf & REQ_NOWAIT) {
@@ -1367,7 +1360,7 @@ static void raid10_write_request(struct mddev *mddev, struct bio *bio,
 		for (;;) {
 			prepare_to_wait(&conf->wait_barrier,
 					&w, TASK_IDLE);
-			if (!md_cluster_ops->area_resyncing(mddev, WRITE,
+			if (!mddev->cluster_ops->area_resyncing(mddev, WRITE,
 				 bio->bi_iter.bi_sector, bio_end_sector(bio)))
 				break;
 			schedule();
@@ -1631,11 +1624,10 @@ static int raid10_handle_discard(struct mddev *mddev, struct bio *bio)
 	if (test_bit(MD_RECOVERY_RESHAPE, &mddev->recovery))
 		return -EAGAIN;
 
-	if (WARN_ON_ONCE(bio->bi_opf & REQ_NOWAIT)) {
+	if (!wait_barrier(conf, bio->bi_opf & REQ_NOWAIT)) {
 		bio_wouldblock_error(bio);
 		return 0;
 	}
-	wait_barrier(conf, false);
 
 	/*
 	 * Check reshape again to avoid reshape happens after checking
@@ -3716,7 +3708,7 @@ static sector_t raid10_sync_request(struct mddev *mddev, sector_t sector_nr,
 			conf->cluster_sync_low = mddev->curr_resync_completed;
 			raid10_set_cluster_sync_high(conf);
 			/* Send resync message */
-			md_cluster_ops->resync_info_update(mddev,
+			mddev->cluster_ops->resync_info_update(mddev,
 						conf->cluster_sync_low,
 						conf->cluster_sync_high);
 		}
@@ -3749,7 +3741,7 @@ static sector_t raid10_sync_request(struct mddev *mddev, sector_t sector_nr,
 		}
 		if (broadcast_msg) {
 			raid10_set_cluster_sync_high(conf);
-			md_cluster_ops->resync_info_update(mddev,
+			mddev->cluster_ops->resync_info_update(mddev,
 						conf->cluster_sync_low,
 						conf->cluster_sync_high);
 		}
@@ -4543,7 +4535,7 @@ static int raid10_start_reshape(struct mddev *mddev)
 		if (ret)
 			goto abort;
 
-		ret = md_cluster_ops->resize_bitmaps(mddev, newsize, oldsize);
+		ret = mddev->cluster_ops->resize_bitmaps(mddev, newsize, oldsize);
 		if (ret) {
 			mddev->bitmap_ops->resize(mddev, oldsize, 0, false);
 			goto abort;
@@ -4834,7 +4826,7 @@ read_more:
 				conf->cluster_sync_low = sb_reshape_pos;
 		}
 
-		md_cluster_ops->resync_info_update(mddev, conf->cluster_sync_low,
+		mddev->cluster_ops->resync_info_update(mddev, conf->cluster_sync_low,
 							  conf->cluster_sync_high);
 	}
 
@@ -4979,7 +4971,7 @@ static void raid10_update_reshape_pos(struct mddev *mddev)
 	struct r10conf *conf = mddev->private;
 	sector_t lo, hi;
 
-	md_cluster_ops->resync_info_get(mddev, &lo, &hi);
+	mddev->cluster_ops->resync_info_get(mddev, &lo, &hi);
 	if (((mddev->reshape_position <= hi) && (mddev->reshape_position >= lo))
 	    || mddev->reshape_position == MaxSector)
 		conf->reshape_progress = mddev->reshape_position;
@@ -5125,9 +5117,13 @@ static void raid10_finish_reshape(struct mddev *mddev)
 
 static struct md_personality raid10_personality =
 {
-	.name		= "raid10",
-	.level		= 10,
-	.owner		= THIS_MODULE,
+	.head = {
+		.type	= MD_PERSONALITY,
+		.id	= ID_RAID10,
+		.name	= "raid10",
+		.owner	= THIS_MODULE,
+	},
+
 	.make_request	= raid10_make_request,
 	.run		= raid10_run,
 	.free		= raid10_free,
@@ -5147,18 +5143,18 @@ static struct md_personality raid10_personality =
 	.update_reshape_pos = raid10_update_reshape_pos,
 };
 
-static int __init raid_init(void)
+static int __init raid10_init(void)
 {
-	return register_md_personality(&raid10_personality);
+	return register_md_submodule(&raid10_personality.head);
 }
 
-static void raid_exit(void)
+static void __exit raid10_exit(void)
 {
-	unregister_md_personality(&raid10_personality);
+	unregister_md_submodule(&raid10_personality.head);
 }
 
-module_init(raid_init);
-module_exit(raid_exit);
+module_init(raid10_init);
+module_exit(raid10_exit);
 MODULE_LICENSE("GPL");
 MODULE_DESCRIPTION("RAID10 (striped mirror) personality for MD");
 MODULE_ALIAS("md-personality-9"); /* RAID10 */
diff --git a/drivers/md/raid5.c b/drivers/md/raid5.c
index 5c79429acc64..6389383166c0 100644
--- a/drivers/md/raid5.c
+++ b/drivers/md/raid5.c
@@ -5858,6 +5858,9 @@ static enum reshape_loc get_reshape_loc(struct mddev *mddev,
 		struct r5conf *conf, sector_t logical_sector)
 {
 	sector_t reshape_progress, reshape_safe;
+
+	if (likely(conf->reshape_progress == MaxSector))
+		return LOC_NO_RESHAPE;
 	/*
 	 * Spinlock is needed as reshape_progress may be
 	 * 64bit on a 32bit platform, and so it might be
@@ -5935,22 +5938,19 @@ static enum stripe_result make_stripe_request(struct mddev *mddev,
 	const int rw = bio_data_dir(bi);
 	enum stripe_result ret;
 	struct stripe_head *sh;
+	enum reshape_loc loc;
 	sector_t new_sector;
 	int previous = 0, flags = 0;
 	int seq, dd_idx;
 
 	seq = read_seqcount_begin(&conf->gen_lock);
-
-	if (unlikely(conf->reshape_progress != MaxSector)) {
-		enum reshape_loc loc = get_reshape_loc(mddev, conf,
-						       logical_sector);
-		if (loc == LOC_INSIDE_RESHAPE) {
-			ret = STRIPE_SCHEDULE_AND_RETRY;
-			goto out;
-		}
-		if (loc == LOC_AHEAD_OF_RESHAPE)
-			previous = 1;
+	loc = get_reshape_loc(mddev, conf, logical_sector);
+	if (loc == LOC_INSIDE_RESHAPE) {
+		ret = STRIPE_SCHEDULE_AND_RETRY;
+		goto out;
 	}
+	if (loc == LOC_AHEAD_OF_RESHAPE)
+		previous = 1;
 
 	new_sector = raid5_compute_sector(conf, logical_sector, previous,
 					  &dd_idx, NULL);
@@ -6127,7 +6127,6 @@ static bool raid5_make_request(struct mddev *mddev, struct bio * bi)
 
 	/* Bail out if conflicts with reshape and REQ_NOWAIT is set */
 	if ((bi->bi_opf & REQ_NOWAIT) &&
-	    (conf->reshape_progress != MaxSector) &&
 	    get_reshape_loc(mddev, conf, logical_sector) == LOC_INSIDE_RESHAPE) {
 		bio_wouldblock_error(bi);
 		if (rw == WRITE)
@@ -8954,9 +8953,13 @@ static void raid5_prepare_suspend(struct mddev *mddev)
 
 static struct md_personality raid6_personality =
 {
-	.name		= "raid6",
-	.level		= 6,
-	.owner		= THIS_MODULE,
+	.head = {
+		.type	= MD_PERSONALITY,
+		.id	= ID_RAID6,
+		.name	= "raid6",
+		.owner	= THIS_MODULE,
+	},
+
 	.make_request	= raid5_make_request,
 	.run		= raid5_run,
 	.start		= raid5_start,
@@ -8980,9 +8983,13 @@ static struct md_personality raid6_personality =
 };
 static struct md_personality raid5_personality =
 {
-	.name		= "raid5",
-	.level		= 5,
-	.owner		= THIS_MODULE,
+	.head = {
+		.type	= MD_PERSONALITY,
+		.id	= ID_RAID5,
+		.name	= "raid5",
+		.owner	= THIS_MODULE,
+	},
+
 	.make_request	= raid5_make_request,
 	.run		= raid5_run,
 	.start		= raid5_start,
@@ -9007,9 +9014,13 @@ static struct md_personality raid5_personality =
 
 static struct md_personality raid4_personality =
 {
-	.name		= "raid4",
-	.level		= 4,
-	.owner		= THIS_MODULE,
+	.head = {
+		.type	= MD_PERSONALITY,
+		.id	= ID_RAID4,
+		.name	= "raid4",
+		.owner	= THIS_MODULE,
+	},
+
 	.make_request	= raid5_make_request,
 	.run		= raid5_run,
 	.start		= raid5_start,
@@ -9045,21 +9056,39 @@ static int __init raid5_init(void)
 				      "md/raid5:prepare",
 				      raid456_cpu_up_prepare,
 				      raid456_cpu_dead);
-	if (ret) {
-		destroy_workqueue(raid5_wq);
-		return ret;
-	}
-	register_md_personality(&raid6_personality);
-	register_md_personality(&raid5_personality);
-	register_md_personality(&raid4_personality);
+	if (ret)
+		goto err_destroy_wq;
+
+	ret = register_md_submodule(&raid6_personality.head);
+	if (ret)
+		goto err_cpuhp_remove;
+
+	ret = register_md_submodule(&raid5_personality.head);
+	if (ret)
+		goto err_unregister_raid6;
+
+	ret = register_md_submodule(&raid4_personality.head);
+	if (ret)
+		goto err_unregister_raid5;
+
 	return 0;
+
+err_unregister_raid5:
+	unregister_md_submodule(&raid5_personality.head);
+err_unregister_raid6:
+	unregister_md_submodule(&raid6_personality.head);
+err_cpuhp_remove:
+	cpuhp_remove_multi_state(CPUHP_MD_RAID5_PREPARE);
+err_destroy_wq:
+	destroy_workqueue(raid5_wq);
+	return ret;
 }
 
-static void raid5_exit(void)
+static void __exit raid5_exit(void)
 {
-	unregister_md_personality(&raid6_personality);
-	unregister_md_personality(&raid5_personality);
-	unregister_md_personality(&raid4_personality);
+	unregister_md_submodule(&raid6_personality.head);
+	unregister_md_submodule(&raid5_personality.head);
+	unregister_md_submodule(&raid4_personality.head);
 	cpuhp_remove_multi_state(CPUHP_MD_RAID5_PREPARE);
 	destroy_workqueue(raid5_wq);
 }