Merge branch 'vfs-7.3.super' into vfs.all

27 files changed, 882 insertions, 306 deletions

diff --git a/block/bdev.c b/block/bdev.c
index 85ce57bd2ae4..797d7f0ef609 100644
--- a/block/bdev.c
+++ b/block/bdev.c
@@ -304,7 +304,12 @@ int bdev_freeze(struct block_device *bdev)
 
 	mutex_lock(&bdev->bd_fsfreeze_mutex);
 
-	if (atomic_inc_return(&bdev->bd_fsfreeze_count) > 1) {
+	/* A device being removed from its filesystem refuses freezes. */
+	if (!atomic_inc_unless_negative(&bdev->bd_fsfreeze_count)) {
+		mutex_unlock(&bdev->bd_fsfreeze_mutex);
+		return -EBUSY;
+	}
+	if (atomic_read(&bdev->bd_fsfreeze_count) > 1) {
 		mutex_unlock(&bdev->bd_fsfreeze_mutex);
 		return 0;
 	}
@@ -340,18 +345,18 @@ int bdev_thaw(struct block_device *bdev)
 
 	mutex_lock(&bdev->bd_fsfreeze_mutex);
 
-	/*
-	 * If this returns < 0 it means that @bd_fsfreeze_count was
-	 * already 0 and no decrement was performed.
-	 */
-	nr_freeze = atomic_dec_if_positive(&bdev->bd_fsfreeze_count);
-	if (nr_freeze < 0)
+	/* <= 0: not frozen (0) or a freeze deny is held (< 0); leave it. */
+	nr_freeze = atomic_read(&bdev->bd_fsfreeze_count);
+	if (nr_freeze <= 0)
 		goto out;
 
 	error = 0;
-	if (nr_freeze > 0)
+	if (nr_freeze > 1) {
+		atomic_dec(&bdev->bd_fsfreeze_count);
 		goto out;
+	}
 
+	/* Keep the count positive across the thaw so a deny is refused. */
 	mutex_lock(&bdev->bd_holder_lock);
 	if (bdev->bd_holder_ops && bdev->bd_holder_ops->thaw) {
 		error = bdev->bd_holder_ops->thaw(bdev);
@@ -360,14 +365,52 @@ int bdev_thaw(struct block_device *bdev)
 		mutex_unlock(&bdev->bd_holder_lock);
 	}
 
-	if (error)
-		atomic_inc(&bdev->bd_fsfreeze_count);
+	if (!error)
+		atomic_dec(&bdev->bd_fsfreeze_count);
 out:
 	mutex_unlock(&bdev->bd_fsfreeze_mutex);
 	return error;
 }
 EXPORT_SYMBOL(bdev_thaw);
 
+/**
+ * bdev_deny_freeze - make a block device unfreezable
+ * @bdev: block device
+ *
+ * Reserve @bdev against bdev_freeze() the way deny_write_access() reserves a
+ * file against writers.  bd_fsfreeze_count is sign-encoded: > 0 counts active
+ * freezes, < 0 counts deniers, so a deny succeeds only while no freeze is in
+ * progress.  While held, bdev_freeze() returns -EBUSY.  Pair with
+ * bdev_allow_freeze().
+ *
+ * A filesystem removing, adding or replacing a member device denies freezes on
+ * it for the duration, so a claim a freeze walk might act on is never torn down
+ * behind the freezer's back.  The deny is device-scoped, not (device,
+ * superblock)-scoped: a device shared by several superblocks is refused for all
+ * of them.  No in-tree filesystem removes a shared claim from a live superblock.
+ *
+ * Return: 0, or -EBUSY if the device is currently frozen.
+ */
+int bdev_deny_freeze(struct block_device *bdev)
+{
+	return atomic_dec_unless_positive(&bdev->bd_fsfreeze_count) ? 0 : -EBUSY;
+}
+EXPORT_SYMBOL_GPL(bdev_deny_freeze);
+
+/**
+ * bdev_allow_freeze - allow freezing a block device again
+ * @bdev: block device
+ *
+ * Undo one bdev_deny_freeze().
+ */
+void bdev_allow_freeze(struct block_device *bdev)
+{
+	/* A deny must be held, i.e. the count must be negative. */
+	WARN_ON_ONCE(atomic_read(&bdev->bd_fsfreeze_count) >= 0);
+	atomic_inc(&bdev->bd_fsfreeze_count);
+}
+EXPORT_SYMBOL_GPL(bdev_allow_freeze);
+
 /*
  * pseudo-fs
  */
@@ -1153,6 +1196,39 @@ put_no_open:
 }
 
 /**
+ * bdev_yield_claim - give up the holder claim on an open block device
+ * @bdev_file: open block device
+ *
+ * Yield the holder and any write access for @bdev_file without closing it, so
+ * the caller can still act on the device - e.g. bdev_allow_freeze() it - before
+ * the final bdev_fput().  bdev_fput() yields too, so calling it afterwards is
+ * safe.
+ */
+void bdev_yield_claim(struct file *bdev_file)
+{
+	struct block_device *bdev;
+	struct gendisk *disk;
+
+	if (!bdev_file->private_data)
+		return;
+
+	bdev = file_bdev(bdev_file);
+	disk = bdev->bd_disk;
+
+	mutex_lock(&disk->open_mutex);
+	bdev_yield_write_access(bdev_file);
+	bd_yield_claim(bdev_file);
+	/*
+	 * Tell release we already gave up our hold on the
+	 * device and if write restrictions are available that
+	 * we already gave up write access to the device.
+	 */
+	bdev_file->private_data = BDEV_I(bdev_file->f_mapping->host);
+	mutex_unlock(&disk->open_mutex);
+}
+EXPORT_SYMBOL_GPL(bdev_yield_claim);
+
+/**
  * bdev_fput - yield claim to the block device and put the file
  * @bdev_file: open block device
  *
@@ -1165,22 +1241,7 @@ void bdev_fput(struct file *bdev_file)
 	if (WARN_ON_ONCE(bdev_file->f_op != &def_blk_fops))
 		return;
 
-	if (bdev_file->private_data) {
-		struct block_device *bdev = file_bdev(bdev_file);
-		struct gendisk *disk = bdev->bd_disk;
-
-		mutex_lock(&disk->open_mutex);
-		bdev_yield_write_access(bdev_file);
-		bd_yield_claim(bdev_file);
-		/*
-		 * Tell release we already gave up our hold on the
-		 * device and if write restrictions are available that
-		 * we already gave up write access to the device.
-		 */
-		bdev_file->private_data = BDEV_I(bdev_file->f_mapping->host);
-		mutex_unlock(&disk->open_mutex);
-	}
-
+	bdev_yield_claim(bdev_file);
 	fput(bdev_file);
 }
 EXPORT_SYMBOL(bdev_fput);
@@ -1217,6 +1278,18 @@ int lookup_bdev(const char *pathname, dev_t *dev)
 	if (!may_open_dev(&path))
 		goto out_path_put;
 
+	/*
+	 * Reject a block device inode with i_rdev == 0.  A dev_t of 0 is
+	 * never valid for a block device: no real block device driver
+	 * registers major 0.  Fake block device inodes (e.g. fuse with
+	 * rootmode=S_IFBLK) can expose i_rdev == 0, and letting that
+	 * propagate would confuse superblock lookup and trigger warnings
+	 * in the device-to-superblock table (super_dev_register).
+	 */
+	error = -ENODEV;
+	if (!inode->i_rdev)
+		goto out_path_put;
+
 	*dev = inode->i_rdev;
 	error = 0;
 out_path_put:
diff --git a/fs/btrfs/dev-replace.c b/fs/btrfs/dev-replace.c
index 318ddb790429..dc0834f920c3 100644
--- a/fs/btrfs/dev-replace.c
+++ b/fs/btrfs/dev-replace.c
@@ -247,8 +247,8 @@ static int btrfs_init_dev_replace_tgtdev(struct btrfs_fs_info *fs_info,
 		return -EINVAL;
 	}
 
-	bdev_file = bdev_file_open_by_path(device_path, BLK_OPEN_WRITE,
-					   fs_info->sb, &fs_holder_ops);
+	/* Unfreezable for the whole replace; see btrfs_dev_replace_start(). */
+	bdev_file = btrfs_open_device_deny_freeze(device_path, fs_info->sb);
 	if (IS_ERR(bdev_file)) {
 		btrfs_err(fs_info, "target device %s is invalid!", device_path);
 		return PTR_ERR(bdev_file);
@@ -327,7 +327,8 @@ static int btrfs_init_dev_replace_tgtdev(struct btrfs_fs_info *fs_info,
 	return 0;
 
 error:
-	bdev_fput(bdev_file);
+	/* Undo the open-time freeze deny. */
+	btrfs_release_device_allow_freeze(bdev_file);
 	return ret;
 }
 
@@ -624,6 +625,15 @@ static int btrfs_dev_replace_start(struct btrfs_fs_info *fs_info,
 	if (ret)
 		return ret;
 
+	/* Deny the source before mark, so every 'leave' unwinds both denied. */
+	if (src_device->bdev) {
+		ret = bdev_deny_freeze(src_device->bdev);
+		if (ret) {
+			btrfs_destroy_dev_replace_tgtdev(tgt_device, true);
+			return ret;
+		}
+	}
+
 	ret = mark_block_group_to_copy(fs_info, src_device);
 	if (ret)
 		return ret;
@@ -708,7 +718,9 @@ static int btrfs_dev_replace_start(struct btrfs_fs_info *fs_info,
 	return ret;
 
 leave:
-	btrfs_destroy_dev_replace_tgtdev(tgt_device);
+	if (src_device->bdev)
+		bdev_allow_freeze(src_device->bdev);
+	btrfs_destroy_dev_replace_tgtdev(tgt_device, true);
 	return ret;
 }
 
@@ -889,6 +901,7 @@ static int btrfs_dev_replace_finishing(struct btrfs_fs_info *fs_info,
 	 */
 	ret = btrfs_start_delalloc_roots(fs_info, LONG_MAX, false);
 	if (ret) {
+		/* Stays started/resumable; keep both denied. */
 		mutex_unlock(&dev_replace->lock_finishing_cancel_unmount);
 		return ret;
 	}
@@ -902,6 +915,7 @@ static int btrfs_dev_replace_finishing(struct btrfs_fs_info *fs_info,
 	while (1) {
 		trans = btrfs_start_transaction(root, 0);
 		if (IS_ERR(trans)) {
+			/* Stays started/resumable; keep both denied. */
 			mutex_unlock(&dev_replace->lock_finishing_cancel_unmount);
 			return PTR_ERR(trans);
 		}
@@ -954,7 +968,10 @@ error:
 		mutex_unlock(&fs_devices->device_list_mutex);
 		btrfs_rm_dev_replace_blocked(fs_info);
 		if (tgt_device)
-			btrfs_destroy_dev_replace_tgtdev(tgt_device);
+			btrfs_destroy_dev_replace_tgtdev(tgt_device, true);
+		/* The source stays a member; re-allow freezing it. */
+		if (src_device->bdev)
+			bdev_allow_freeze(src_device->bdev);
 		btrfs_rm_dev_replace_unblocked(fs_info);
 		mutex_unlock(&dev_replace->lock_finishing_cancel_unmount);
 
@@ -1027,6 +1044,8 @@ error:
 
 	mutex_unlock(&dev_replace->lock_finishing_cancel_unmount);
 
+	/* The target is now a member; the source is freed (allow + release). */
+	bdev_allow_freeze(tgt_device->bdev);
 	btrfs_rm_dev_replace_free_srcdev(src_device);
 
 	return 0;
@@ -1155,8 +1174,9 @@ int btrfs_dev_replace_cancel(struct btrfs_fs_info *fs_info)
 			btrfs_dev_name(src_device), src_device->devid,
 			btrfs_dev_name(tgt_device));
 
+		/* A suspended replace never re-denied freezing; do not allow. */
 		if (tgt_device)
-			btrfs_destroy_dev_replace_tgtdev(tgt_device);
+			btrfs_destroy_dev_replace_tgtdev(tgt_device, false);
 		break;
 	default:
 		up_write(&dev_replace->rwsem);
@@ -1186,6 +1206,11 @@ void btrfs_dev_replace_suspend_for_unmount(struct btrfs_fs_info *fs_info)
 		dev_replace->time_stopped = ktime_get_real_seconds();
 		dev_replace->item_needs_writeback = 1;
 		btrfs_info(fs_info, "suspending dev_replace for unmount");
+		/* Reopened freezable next mount; resume re-denies. */
+		if (dev_replace->srcdev && dev_replace->srcdev->bdev)
+			bdev_allow_freeze(dev_replace->srcdev->bdev);
+		if (dev_replace->tgtdev && dev_replace->tgtdev->bdev)
+			bdev_allow_freeze(dev_replace->tgtdev->bdev);
 		break;
 	}
 
@@ -1198,6 +1223,7 @@ int btrfs_resume_dev_replace_async(struct btrfs_fs_info *fs_info)
 {
 	struct task_struct *task;
 	struct btrfs_dev_replace *dev_replace = &fs_info->dev_replace;
+	int ret = 0;
 
 	down_write(&dev_replace->rwsem);
 
@@ -1241,8 +1267,33 @@ int btrfs_resume_dev_replace_async(struct btrfs_fs_info *fs_info)
 		return 0;
 	}
 
+	/* Re-deny for the resumed replace; stay suspended if frozen now. */
+	if (dev_replace->srcdev->bdev &&
+	    bdev_deny_freeze(dev_replace->srcdev->bdev))
+		goto suspend;
+	if (bdev_deny_freeze(dev_replace->tgtdev->bdev)) {
+		if (dev_replace->srcdev->bdev)
+			bdev_allow_freeze(dev_replace->srcdev->bdev);
+		goto suspend;
+	}
+
 	task = kthread_run(btrfs_dev_replace_kthread, fs_info, "btrfs-devrepl");
-	return PTR_ERR_OR_ZERO(task);
+	if (IS_ERR(task)) {
+		bdev_allow_freeze(dev_replace->tgtdev->bdev);
+		if (dev_replace->srcdev->bdev)
+			bdev_allow_freeze(dev_replace->srcdev->bdev);
+		/* Undo the deny and suspend, but still fail the mount. */
+		ret = PTR_ERR(task);
+		goto suspend;
+	}
+	return 0;
+
+suspend:
+	btrfs_exclop_finish(fs_info);
+	down_write(&dev_replace->rwsem);
+	dev_replace->replace_state = BTRFS_IOCTL_DEV_REPLACE_STATE_SUSPENDED;
+	up_write(&dev_replace->rwsem);
+	return ret;
 }
 
 static int btrfs_dev_replace_kthread(void *data)
diff --git a/fs/btrfs/ioctl.c b/fs/btrfs/ioctl.c
index baa645e98812..c4e995661bcf 100644
--- a/fs/btrfs/ioctl.c
+++ b/fs/btrfs/ioctl.c
@@ -2629,7 +2629,7 @@ static long btrfs_ioctl_rm_dev_v2(struct file *file, void __user *arg)
 err_drop:
 	mnt_drop_write_file(file);
 	if (bdev_file)
-		bdev_fput(bdev_file);
+		btrfs_release_device_allow_freeze(bdev_file);
 out:
 	btrfs_put_dev_args_from_path(&args);
 	return ret;
@@ -2679,7 +2679,7 @@ static long btrfs_ioctl_rm_dev(struct file *file, void __user *arg)
 
 	mnt_drop_write_file(file);
 	if (bdev_file)
-		bdev_fput(bdev_file);
+		btrfs_release_device_allow_freeze(bdev_file);
 out:
 	btrfs_put_dev_args_from_path(&args);
 	return ret;
diff --git a/fs/btrfs/volumes.c b/fs/btrfs/volumes.c
index dc5f4a122d55..e68ce323bb06 100644
--- a/fs/btrfs/volumes.c
+++ b/fs/btrfs/volumes.c
@@ -481,7 +481,12 @@ btrfs_get_bdev_and_sb(const char *device_path, blk_mode_t flags, void *holder,
 	struct block_device *bdev;
 	int ret;
 
-	*bdev_file = bdev_file_open_by_path(device_path, flags, holder, &fs_holder_ops);
+	if (holder)
+		*bdev_file = fs_bdev_file_open_by_path(device_path, flags,
+						       holder, holder);
+	else
+		*bdev_file = bdev_file_open_by_path(device_path, flags, NULL,
+						    NULL);
 
 	if (IS_ERR(*bdev_file)) {
 		ret = PTR_ERR(*bdev_file);
@@ -496,7 +501,7 @@ btrfs_get_bdev_and_sb(const char *device_path, blk_mode_t flags, void *holder,
 	if (holder) {
 		ret = set_blocksize(*bdev_file, BTRFS_BDEV_BLOCKSIZE);
 		if (ret) {
-			bdev_fput(*bdev_file);
+			fs_bdev_file_release(*bdev_file, holder);
 			goto error;
 		}
 	}
@@ -504,7 +509,10 @@ btrfs_get_bdev_and_sb(const char *device_path, blk_mode_t flags, void *holder,
 	*disk_super = btrfs_read_disk_super(bdev, 0, false);
 	if (IS_ERR(*disk_super)) {
 		ret = PTR_ERR(*disk_super);
-		bdev_fput(*bdev_file);
+		if (holder)
+			fs_bdev_file_release(*bdev_file, holder);
+		else
+			bdev_fput(*bdev_file);
 		goto error;
 	}
 
@@ -728,7 +736,7 @@ static int btrfs_open_one_device(struct btrfs_fs_devices *fs_devices,
 
 error_free_page:
 	btrfs_release_disk_super(disk_super);
-	bdev_fput(bdev_file);
+	fs_bdev_file_release(bdev_file, holder);
 
 	return -EINVAL;
 }
@@ -1088,7 +1096,7 @@ static void __btrfs_free_extra_devids(struct btrfs_fs_devices *fs_devices,
 			continue;
 
 		if (device->bdev_file) {
-			bdev_fput(device->bdev_file);
+			fs_bdev_file_release(device->bdev_file, device->bdev_file->private_data);
 			device->bdev = NULL;
 			device->bdev_file = NULL;
 			fs_devices->open_devices--;
@@ -1125,7 +1133,18 @@ void btrfs_free_extra_devids(struct btrfs_fs_devices *fs_devices)
 	mutex_unlock(&uuid_mutex);
 }
 
-static void btrfs_close_bdev(struct btrfs_device *device)
+/* Release a device that was made unfreezable for a membership change. */
+void btrfs_release_device_allow_freeze(struct file *bdev_file)
+{
+	struct super_block *sb = bdev_file->private_data;
+
+	/* Unregister before re-allowing (strand-safe); file still open (UAF-safe). */
+	fs_bdev_unregister(bdev_file, sb);
+	bdev_allow_freeze(file_bdev(bdev_file));
+	bdev_fput(bdev_file);
+}
+
+static void btrfs_close_bdev(struct btrfs_device *device, bool allow_freeze)
 {
 	if (!device->bdev)
 		return;
@@ -1135,7 +1154,12 @@ static void btrfs_close_bdev(struct btrfs_device *device)
 		invalidate_bdev(device->bdev);
 	}
 
-	bdev_fput(device->bdev_file);
+	/* @allow_freeze undoes a replace-time deny; unmount-close was never denied. */
+	if (allow_freeze)
+		btrfs_release_device_allow_freeze(device->bdev_file);
+	else
+		fs_bdev_file_release(device->bdev_file,
+				     device->bdev_file->private_data);
 }
 
 static void btrfs_close_one_device(struct btrfs_device *device)
@@ -1156,7 +1180,7 @@ static void btrfs_close_one_device(struct btrfs_device *device)
 		fs_devices->missing_devices--;
 	}
 
-	btrfs_close_bdev(device);
+	btrfs_close_bdev(device, false);
 	if (device->bdev) {
 		fs_devices->open_devices--;
 		device->bdev = NULL;
@@ -2382,6 +2406,13 @@ int btrfs_rm_device(struct btrfs_fs_info *fs_info,
 	    fs_info->fs_devices->rw_devices == 1)
 		return BTRFS_ERROR_DEV_ONLY_WRITABLE;
 
+	/* Removal and freezing are mutually exclusive; refuse if frozen now. */
+	if (device->bdev) {
+		ret = bdev_deny_freeze(device->bdev);
+		if (ret)
+			return ret;
+	}
+
 	if (test_bit(BTRFS_DEV_STATE_WRITEABLE, &device->dev_state)) {
 		mutex_lock(&fs_info->chunk_mutex);
 		list_del_init(&device->dev_alloc_list);
@@ -2408,6 +2439,8 @@ int btrfs_rm_device(struct btrfs_fs_info *fs_info,
 			   device->devid, ret);
 		btrfs_abort_transaction(trans, ret);
 		btrfs_end_transaction(trans);
+		if (device->bdev)
+			bdev_allow_freeze(device->bdev);
 		return ret;
 	}
 
@@ -2499,6 +2532,8 @@ int btrfs_rm_device(struct btrfs_fs_info *fs_info,
 	return btrfs_commit_transaction(trans);
 
 error_undo:
+	if (device->bdev)
+		bdev_allow_freeze(device->bdev);
 	if (test_bit(BTRFS_DEV_STATE_WRITEABLE, &device->dev_state)) {
 		mutex_lock(&fs_info->chunk_mutex);
 		list_add(&device->dev_alloc_list,
@@ -2543,7 +2578,8 @@ void btrfs_rm_dev_replace_free_srcdev(struct btrfs_device *srcdev)
 
 	mutex_lock(&uuid_mutex);
 
-	btrfs_close_bdev(srcdev);
+	/* The source was made unfreezable for the replace; undo it. */
+	btrfs_close_bdev(srcdev, true);
 	synchronize_rcu();
 	btrfs_free_device(srcdev);
 
@@ -2564,7 +2600,8 @@ void btrfs_rm_dev_replace_free_srcdev(struct btrfs_device *srcdev)
 	mutex_unlock(&uuid_mutex);
 }
 
-void btrfs_destroy_dev_replace_tgtdev(struct btrfs_device *tgtdev)
+void btrfs_destroy_dev_replace_tgtdev(struct btrfs_device *tgtdev,
+				      bool allow_freeze)
 {
 	struct btrfs_fs_devices *fs_devices = tgtdev->fs_info->fs_devices;
 
@@ -2585,7 +2622,7 @@ void btrfs_destroy_dev_replace_tgtdev(struct btrfs_device *tgtdev)
 
 	btrfs_scratch_superblocks(tgtdev->fs_info, tgtdev);
 
-	btrfs_close_bdev(tgtdev);
+	btrfs_close_bdev(tgtdev, allow_freeze);
 	synchronize_rcu();
 	btrfs_free_device(tgtdev);
 }
@@ -2854,6 +2891,37 @@ next_slot:
 	return 0;
 }
 
+/*
+ * Open @path for @sb with freezing denied before the holder claim is published,
+ * so a racing bdev_freeze() can never reach a claim a device add or replace may
+ * still abort.  The deny is taken on a throwaway non-holder probe open, then the
+ * holder is opened by the probe's dev_t.  Balanced by the caller.
+ */
+struct file *btrfs_open_device_deny_freeze(const char *path,
+					   struct super_block *sb)
+{
+	struct file *probe_file, *bdev_file;
+	int ret;
+
+	/* WRITE so bdev_file_open_by_path() rejects a read-only device. */
+	probe_file = bdev_file_open_by_path(path, BLK_OPEN_WRITE, NULL, NULL);
+	if (IS_ERR(probe_file))
+		return probe_file;
+
+	ret = bdev_deny_freeze(file_bdev(probe_file));
+	if (ret) {
+		bdev_fput(probe_file);
+		return ERR_PTR(ret);
+	}
+
+	bdev_file = fs_bdev_file_open_by_dev(file_bdev(probe_file)->bd_dev,
+					     BLK_OPEN_WRITE, sb, sb);
+	if (IS_ERR(bdev_file))
+		bdev_allow_freeze(file_bdev(probe_file));
+	bdev_fput(probe_file);
+	return bdev_file;
+}
+
 int btrfs_init_new_device(struct btrfs_fs_info *fs_info, const char *device_path)
 {
 	struct btrfs_root *root = fs_info->dev_root;
@@ -2872,8 +2940,8 @@ int btrfs_init_new_device(struct btrfs_fs_info *fs_info, const char *device_path
 	if (sb_rdonly(sb) && !fs_devices->seeding)
 		return -EROFS;
 
-	bdev_file = bdev_file_open_by_path(device_path, BLK_OPEN_WRITE,
-					   fs_info->sb, &fs_holder_ops);
+	/* Forbid freezing until the device is a committed member (or unwound). */
+	bdev_file = btrfs_open_device_deny_freeze(device_path, fs_info->sb);
 	if (IS_ERR(bdev_file))
 		return PTR_ERR(bdev_file);
 
@@ -3044,8 +3112,10 @@ int btrfs_init_new_device(struct btrfs_fs_info *fs_info, const char *device_path
 		up_write(&sb->s_umount);
 		locked = false;
 
-		if (ret) /* transaction commit */
+		if (ret) { /* transaction commit */
+			bdev_allow_freeze(file_bdev(bdev_file));
 			return ret;
+		}
 
 		ret = btrfs_relocate_sys_chunks(fs_info);
 		if (ret < 0)
@@ -3053,8 +3123,10 @@ int btrfs_init_new_device(struct btrfs_fs_info *fs_info, const char *device_path
 				    "Failed to relocate sys chunks after device initialization. This can be fixed using the \"btrfs balance\" command.");
 		trans = btrfs_attach_transaction(root);
 		if (IS_ERR(trans)) {
-			if (PTR_ERR(trans) == -ENOENT)
+			if (PTR_ERR(trans) == -ENOENT) {
+				bdev_allow_freeze(file_bdev(bdev_file));
 				return 0;
+			}
 			ret = PTR_ERR(trans);
 			trans = NULL;
 			goto error_sysfs;
@@ -3074,6 +3146,7 @@ int btrfs_init_new_device(struct btrfs_fs_info *fs_info, const char *device_path
 	/* Update ctime/mtime for blkid or udev */
 	update_dev_time(device_path);
 
+	bdev_allow_freeze(file_bdev(bdev_file));
 	return ret;
 
 error_sysfs:
@@ -3103,7 +3176,7 @@ error_free_zone:
 error_free_device:
 	btrfs_free_device(device);
 error:
-	bdev_fput(bdev_file);
+	btrfs_release_device_allow_freeze(bdev_file);
 	if (locked) {
 		mutex_unlock(&uuid_mutex);
 		up_write(&sb->s_umount);
diff --git a/fs/btrfs/volumes.h b/fs/btrfs/volumes.h
index 63be45c3298c..df2c671ab6fa 100644
--- a/fs/btrfs/volumes.h
+++ b/fs/btrfs/volumes.h
@@ -744,6 +744,7 @@ int btrfs_open_devices(struct btrfs_fs_devices *fs_devices,
 struct btrfs_device *btrfs_scan_one_device(const char *path, bool mount_arg_dev);
 int btrfs_forget_devices(dev_t devt);
 void btrfs_close_devices(struct btrfs_fs_devices *fs_devices);
+void btrfs_release_device_allow_freeze(struct file *bdev_file);
 void btrfs_free_extra_devids(struct btrfs_fs_devices *fs_devices);
 void btrfs_assign_next_active_device(struct btrfs_device *device,
 				     struct btrfs_device *this_dev);
@@ -768,6 +769,8 @@ struct btrfs_device *btrfs_find_device(const struct btrfs_fs_devices *fs_devices
 				       const struct btrfs_dev_lookup_args *args);
 int btrfs_shrink_device(struct btrfs_device *device, u64 new_size);
 int btrfs_init_new_device(struct btrfs_fs_info *fs_info, const char *path);
+struct file *btrfs_open_device_deny_freeze(const char *path,
+					   struct super_block *sb);
 int btrfs_balance(struct btrfs_fs_info *fs_info,
 		  struct btrfs_balance_control *bctl,
 		  struct btrfs_ioctl_balance_args *bargs);
@@ -788,7 +791,8 @@ int btrfs_init_writeback_bio_size(struct btrfs_fs_info *fs_info);
 int btrfs_run_dev_stats(struct btrfs_trans_handle *trans);
 void btrfs_rm_dev_replace_remove_srcdev(struct btrfs_device *srcdev);
 void btrfs_rm_dev_replace_free_srcdev(struct btrfs_device *srcdev);
-void btrfs_destroy_dev_replace_tgtdev(struct btrfs_device *tgtdev);
+void btrfs_destroy_dev_replace_tgtdev(struct btrfs_device *tgtdev,
+				      bool allow_freeze);
 unsigned long btrfs_full_stripe_len(struct btrfs_fs_info *fs_info,
 				    u64 logical);
 u64 btrfs_calc_stripe_length(const struct btrfs_chunk_map *map);
diff --git a/fs/cramfs/inode.c b/fs/cramfs/inode.c
index 4edbfccd0bbe..d4cd03f4f60d 100644
--- a/fs/cramfs/inode.c
+++ b/fs/cramfs/inode.c
@@ -504,7 +504,7 @@ static void cramfs_kill_sb(struct super_block *sb)
 		sb->s_mtd = NULL;
 	} else if (IS_ENABLED(CONFIG_CRAMFS_BLOCKDEV) && sb->s_bdev) {
 		sync_blockdev(sb->s_bdev);
-		bdev_fput(sb->s_bdev_file);
+		fs_bdev_file_release(sb->s_bdev_file, sb);
 	}
 	kfree(sbi);
 }
diff --git a/fs/erofs/super.c b/fs/erofs/super.c
index 9d8f862f309f..770808685934 100644
--- a/fs/erofs/super.c
+++ b/fs/erofs/super.c
@@ -147,8 +147,8 @@ static int erofs_init_device(struct erofs_buf *buf, struct super_block *sb,
 	if (!sbi->devs->flatdev) {
 		file = erofs_is_fileio_mode(sbi) ?
 				filp_open(dif->path, O_RDONLY | O_LARGEFILE, 0) :
-				bdev_file_open_by_path(dif->path,
-						BLK_OPEN_READ, sb->s_type, NULL);
+				fs_bdev_file_open_by_path(dif->path,
+						BLK_OPEN_READ, sb->s_type, sb);
 		if (IS_ERR(file)) {
 			if (file == ERR_PTR(-ENOTBLK))
 				return -EINVAL;
@@ -790,28 +790,34 @@ static int erofs_fc_reconfigure(struct fs_context *fc)
 
 static int erofs_release_device_info(int id, void *ptr, void *data)
 {
+	struct super_block *sb = data;
 	struct erofs_device_info *dif = ptr;
 
 	fs_put_dax(dif->dax_dev, NULL);
-	if (dif->file)
-		fput(dif->file);
+	if (dif->file) {
+		if (S_ISBLK(file_inode(dif->file)->i_mode))
+			fs_bdev_file_release(dif->file, sb);
+		else
+			fput(dif->file);
+	}
 	kfree(dif->path);
 	kfree(dif);
 	return 0;
 }
 
-static void erofs_free_dev_context(struct erofs_dev_context *devs)
+static void erofs_free_dev_context(struct erofs_dev_context *devs,
+				   struct super_block *sb)
 {
 	if (!devs)
 		return;
-	idr_for_each(&devs->tree, &erofs_release_device_info, NULL);
+	idr_for_each(&devs->tree, &erofs_release_device_info, sb);
 	idr_destroy(&devs->tree);
 	kfree(devs);
 }
 
-static void erofs_sb_free(struct erofs_sb_info *sbi)
+static void erofs_sb_free(struct erofs_sb_info *sbi, struct super_block *sb)
 {
-	erofs_free_dev_context(sbi->devs);
+	erofs_free_dev_context(sbi->devs, sb);
 	kfree_sensitive(sbi->domain_id);
 	if (sbi->dif0.file)
 		fput(sbi->dif0.file);
@@ -823,8 +829,13 @@ static void erofs_fc_free(struct fs_context *fc)
 {
 	struct erofs_sb_info *sbi = fc->s_fs_info;
 
-	if (sbi) /* free here if an error occurs before transferring to sb */
-		erofs_sb_free(sbi);
+	/*
+	 * Freed here only if an error occurs before the sb is set up; at that
+	 * point no block-backed device has been claimed (that happens in
+	 * fill_super), so the NULL sb never reaches fs_bdev_file_release().
+	 */
+	if (sbi)
+		erofs_sb_free(sbi, NULL);
 }
 
 static const struct fs_context_operations erofs_context_ops = {
@@ -878,7 +889,7 @@ static void erofs_kill_sb(struct super_block *sb)
 		kill_block_super(sb);
 	erofs_drop_internal_inodes(sbi);
 	fs_put_dax(sbi->dif0.dax_dev, NULL);
-	erofs_sb_free(sbi);
+	erofs_sb_free(sbi, sb);
 	sb->s_fs_info = NULL;
 }
 
@@ -890,7 +901,7 @@ static void erofs_put_super(struct super_block *sb)
 	erofs_shrinker_unregister(sb);
 	erofs_xattr_prefixes_cleanup(sb);
 	erofs_drop_internal_inodes(sbi);
-	erofs_free_dev_context(sbi->devs);
+	erofs_free_dev_context(sbi->devs, sb);
 	sbi->devs = NULL;
 }
 
diff --git a/fs/ext4/extents-test.c b/fs/ext4/extents-test.c
index bd7795a82607..c3836ecb89f9 100644
--- a/fs/ext4/extents-test.c
+++ b/fs/ext4/extents-test.c
@@ -126,11 +126,6 @@ struct kunit_ext_test_param {
 	struct kunit_ext_data_state exp_data_state[3];
 };
 
-static void ext_kill_sb(struct super_block *sb)
-{
-	generic_shutdown_super(sb);
-}
-
 static int ext_init_fs_context(struct fs_context *fc)
 {
 	return 0;
@@ -138,13 +133,13 @@ static int ext_init_fs_context(struct fs_context *fc)
 
 static int ext_set(struct super_block *sb, struct fs_context *fc)
 {
-	return 0;
+	return set_anon_super_fc(sb, fc);
 }
 
 static struct file_system_type ext_fs_type = {
 	.name		 = "extents test",
 	.init_fs_context = ext_init_fs_context,
-	.kill_sb	 = ext_kill_sb,
+	.kill_sb	 = kill_anon_super,
 };
 
 static void extents_kunit_exit(struct kunit *test)
diff --git a/fs/ext4/mballoc-test.c b/fs/ext4/mballoc-test.c
index 0424b8b0b4c3..d31780075c21 100644
--- a/fs/ext4/mballoc-test.c
+++ b/fs/ext4/mballoc-test.c
@@ -59,11 +59,6 @@ static const struct super_operations mbt_sops = {
 	.free_inode	= mbt_free_inode,
 };
 
-static void mbt_kill_sb(struct super_block *sb)
-{
-	generic_shutdown_super(sb);
-}
-
 static int mbt_init_fs_context(struct fs_context *fc)
 {
 	return 0;
@@ -72,7 +67,7 @@ static int mbt_init_fs_context(struct fs_context *fc)
 static struct file_system_type mbt_fs_type = {
 	.name			= "mballoc test",
 	.init_fs_context	= mbt_init_fs_context,
-	.kill_sb		= mbt_kill_sb,
+	.kill_sb		= kill_anon_super,
 };
 
 static int mbt_mb_init(struct super_block *sb)
@@ -136,7 +131,7 @@ static void mbt_mb_release(struct super_block *sb)
 
 static int mbt_set(struct super_block *sb, struct fs_context *fc)
 {
-	return 0;
+	return set_anon_super_fc(sb, fc);
 }
 
 static struct super_block *mbt_ext4_alloc_super_block(void)
diff --git a/fs/ext4/super.c b/fs/ext4/super.c
index 245f67d10ded..467aa44109bc 100644
--- a/fs/ext4/super.c
+++ b/fs/ext4/super.c
@@ -5797,7 +5797,7 @@ failed_mount:
 	brelse(sbi->s_sbh);
 	if (sbi->s_journal_bdev_file) {
 		invalidate_bdev(file_bdev(sbi->s_journal_bdev_file));
-		bdev_fput(sbi->s_journal_bdev_file);
+		fs_bdev_file_release(sbi->s_journal_bdev_file, sb);
 	}
 out_fail:
 	invalidate_bdev(sb->s_bdev);
@@ -5981,9 +5981,9 @@ static struct file *ext4_get_journal_blkdev(struct super_block *sb,
 	struct ext4_super_block *es;
 	int errno;
 
-	bdev_file = bdev_file_open_by_dev(j_dev,
+	bdev_file = fs_bdev_file_open_by_dev(j_dev,
 		BLK_OPEN_READ | BLK_OPEN_WRITE | BLK_OPEN_RESTRICT_WRITES,
-		sb, &fs_holder_ops);
+		sb, sb);
 	if (IS_ERR(bdev_file)) {
 		ext4_msg(sb, KERN_ERR,
 			 "failed to open journal device unknown-block(%u,%u) %pe",
@@ -6043,7 +6043,7 @@ static struct file *ext4_get_journal_blkdev(struct super_block *sb,
 out_bh:
 	brelse(bh);
 out_bdev:
-	bdev_fput(bdev_file);
+	fs_bdev_file_release(bdev_file, sb);
 	return ERR_PTR(errno);
 }
 
@@ -6082,7 +6082,7 @@ static journal_t *ext4_open_dev_journal(struct super_block *sb,
 out_journal:
 	ext4_journal_destroy(EXT4_SB(sb), journal);
 out_bdev:
-	bdev_fput(bdev_file);
+	fs_bdev_file_release(bdev_file, sb);
 	return ERR_PTR(errno);
 }
 
@@ -7499,7 +7499,7 @@ static void ext4_kill_sb(struct super_block *sb)
 	kill_block_super(sb);
 
 	if (bdev_file)
-		bdev_fput(bdev_file);
+		fs_bdev_file_release(bdev_file, sb);
 }
 
 static struct file_system_type ext4_fs_type = {
diff --git a/fs/f2fs/super.c b/fs/f2fs/super.c
index 9760e4efeffe..a7df36986616 100644
--- a/fs/f2fs/super.c
+++ b/fs/f2fs/super.c
@@ -1971,7 +1971,7 @@ static void destroy_device_list(struct f2fs_sb_info *sbi)
 
 	for (i = 0; i < sbi->s_ndevs; i++) {
 		if (i > 0)
-			bdev_fput(FDEV(i).bdev_file);
+			fs_bdev_file_release(FDEV(i).bdev_file, sbi->sb);
 #ifdef CONFIG_BLK_DEV_ZONED
 		kvfree(FDEV(i).blkz_seq);
 #endif
@@ -4901,8 +4901,8 @@ static int f2fs_scan_devices(struct f2fs_sb_info *sbi)
 				FDEV(i).end_blk = FDEV(i).start_blk +
 						SEGS_TO_BLKS(sbi,
 						FDEV(i).total_segments) - 1;
-				FDEV(i).bdev_file = bdev_file_open_by_path(
-					FDEV(i).path, mode, sbi->sb, NULL);
+				FDEV(i).bdev_file = fs_bdev_file_open_by_path(
+					FDEV(i).path, mode, sbi->sb, sbi->sb);
 			}
 		}
 		if (IS_ERR(FDEV(i).bdev_file))
diff --git a/fs/internal.h b/fs/internal.h
index 67aa0444351b..c658c8a5ebd5 100644
--- a/fs/internal.h
+++ b/fs/internal.h
@@ -137,6 +137,7 @@ extern int reconfigure_super(struct fs_context *);
 extern bool super_trylock_shared(struct super_block *sb);
 struct super_block *user_get_super(dev_t, bool excl);
 void put_super(struct super_block *sb);
+void __init super_dev_init(void);
 extern bool mount_capable(struct fs_context *);
 
 /*
diff --git a/fs/namespace.c b/fs/namespace.c
index 3c382c2bdeec..1ecd96c918b3 100644
--- a/fs/namespace.c
+++ b/fs/namespace.c
@@ -6270,6 +6270,8 @@ void __init mnt_init(void)
 				HASH_ZERO,
 				&mp_hash_shift, &mp_hash_mask, 0, 0);
 
+	super_dev_init();
+
 	kernfs_init();
 
 	err = sysfs_init();
diff --git a/fs/ocfs2/super.c b/fs/ocfs2/super.c
index bcac614ff02a..c62e389d4dd6 100644
--- a/fs/ocfs2/super.c
+++ b/fs/ocfs2/super.c
@@ -1882,7 +1882,6 @@ static void ocfs2_dismount_volume(struct super_block *sb, int mnt_err)
 
 	ocfs2_delete_osb(osb);
 	kfree(osb);
-	sb->s_dev = 0;
 	sb->s_fs_info = NULL;
 }
 
diff --git a/fs/romfs/super.c b/fs/romfs/super.c
index 3a836af3ca7e..7f783d6173e8 100644
--- a/fs/romfs/super.c
+++ b/fs/romfs/super.c
@@ -595,7 +595,7 @@ static void romfs_kill_sb(struct super_block *sb)
 #ifdef CONFIG_ROMFS_ON_BLOCK
 	if (sb->s_bdev) {
 		sync_blockdev(sb->s_bdev);
-		bdev_fput(sb->s_bdev_file);
+		fs_bdev_file_release(sb->s_bdev_file, sb);
 	}
 #endif
 }
diff --git a/fs/super.c b/fs/super.c
index 2c52c321885f..05e443173038 100644
--- a/fs/super.c
+++ b/fs/super.c
@@ -25,6 +25,7 @@
 #include <linux/slab.h>
 #include <linux/blkdev.h>
 #include <linux/memcontrol.h>
+#include <linux/rhashtable.h>
 #include <linux/mount.h>
 #include <linux/security.h>
 #include <linux/writeback.h>		/* for the emergency remount stuff */
@@ -103,7 +104,7 @@ static bool super_flags(const struct super_block *sb, unsigned int flags)
  * creation will succeed and SB_BORN is set by vfs_get_tree() or we're
  * woken and we'll see SB_DYING.
  *
- * The caller must have acquired a temporary reference on @sb->s_count.
+ * The caller must have acquired a temporary reference on @sb->s_passive.
  *
  * Return: The function returns true if SB_BORN was set and with
  *         s_umount held. The function returns false if SB_DYING was
@@ -287,6 +288,8 @@ static unsigned long super_cache_count(struct shrinker *shrink,
 	return total_objects;
 }
 
+static struct super_dev *super_dev_alloc(dev_t dev, struct super_block *sb);
+
 static void destroy_super_work(struct work_struct *work)
 {
 	struct super_block *s = container_of(work, struct super_block,
@@ -294,6 +297,8 @@ static void destroy_super_work(struct work_struct *work)
 	fsnotify_sb_free(s);
 	security_sb_free(s);
 	put_user_ns(s->s_user_ns);
+	/* Only an unregistered entry is still owned by the superblock. */
+	kfree(s->s_super_dev);
 	kfree(s->s_subtype);
 	for (int i = 0; i < SB_FREEZE_LEVELS; i++)
 		percpu_free_rwsem(&s->s_writers.rw_sem[i]);
@@ -382,7 +387,7 @@ static struct super_block *alloc_super(struct file_system_type *type, int flags,
 	spin_lock_init(&s->s_inode_wblist_lock);
 	fserror_mount(s);
 
-	s->s_count = 1;
+	refcount_set(&s->s_passive, 1);
 	atomic_set(&s->s_active, 1);
 	mutex_init(&s->s_vfs_rename_mutex);
 	lockdep_set_class(&s->s_vfs_rename_mutex, &type->s_vfs_rename_key);
@@ -407,6 +412,10 @@ static struct super_block *alloc_super(struct file_system_type *type, int flags,
 		goto fail;
 	if (list_lru_init_memcg(&s->s_inode_lru, s->s_shrink))
 		goto fail;
+	s->s_super_dev = super_dev_alloc(0, s);
+	if (!s->s_super_dev)
+		goto fail;
+
 	s->s_min_writeback_pages = MIN_WRITEBACK_PAGES;
 	return s;
 
@@ -418,12 +427,17 @@ fail:
 /* Superblock refcounting  */
 
 /*
- * Drop a superblock's refcount.  The caller must hold sb_lock.
+ * Drop a superblock's passive reference.  Must be called WITHOUT sb_lock held;
+ * put_super() acquires sb_lock itself when the final reference is dropped.
  */
-static void __put_super(struct super_block *s)
+void put_super(struct super_block *s)
 {
-	if (!--s->s_count) {
+	if (refcount_dec_and_test(&s->s_passive)) {
+
+		spin_lock(&sb_lock);
 		list_del_init(&s->s_list);
+		spin_unlock(&sb_lock);
+
 		WARN_ON(s->s_dentry_lru.node);
 		WARN_ON(s->s_inode_lru.node);
 		WARN_ON(s->s_mounts);
@@ -431,18 +445,109 @@ static void __put_super(struct super_block *s)
 	}
 }
 
-/**
- *	put_super	-	drop a temporary reference to superblock
- *	@sb: superblock in question
- *
- *	Drops a temporary reference, frees superblock if there's no
- *	references left.
- */
-void put_super(struct super_block *sb)
+struct super_dev {
+	dev_t			sd_dev;
+	struct super_block	*sd_sb;
+	refcount_t		sd_ref;
+	struct rhlist_head	sd_node;
+	struct rcu_head		sd_rcu;
+};
+
+static struct rhltable super_dev_table;
+static const struct rhashtable_params super_dev_params = {
+	.key_len	= sizeof(dev_t),
+	.key_offset	= offsetof(struct super_dev, sd_dev),
+	.head_offset	= offsetof(struct super_dev, sd_node),
+};
+
+static struct super_dev *super_dev_alloc(dev_t dev, struct super_block *sb)
 {
-	spin_lock(&sb_lock);
-	__put_super(sb);
-	spin_unlock(&sb_lock);
+	struct super_dev *fsd;
+
+	fsd = kzalloc_obj(*fsd);
+	if (!fsd)
+		return NULL;
+	fsd->sd_dev = dev;
+	fsd->sd_sb = sb;
+	refcount_set(&fsd->sd_ref, 1);
+	return fsd;
+}
+
+static void super_dev_put(struct super_dev *fsd)
+{
+	/* Unlink only once unpinned, so a cursor never resumes from a removed node. */
+	if (fsd && refcount_dec_and_test(&fsd->sd_ref)) {
+		rhltable_remove(&super_dev_table, &fsd->sd_node, super_dev_params);
+		put_super(fsd->sd_sb);
+		kfree_rcu(fsd, sd_rcu);
+	}
+}
+
+void __init super_dev_init(void)
+{
+	if (rhltable_init(&super_dev_table, &super_dev_params))
+		panic("VFS: Cannot initialise super_dev_table\n");
+}
+
+static int super_dev_insert(struct super_dev *fsd)
+{
+	int err;
+
+	err = rhltable_insert(&super_dev_table, &fsd->sd_node, super_dev_params);
+	if (!err)
+		refcount_inc(&fsd->sd_sb->s_passive);
+	return err;
+}
+
+/* Register @sb under @sb->s_dev as the final fallible act of a set callback. */
+static int super_dev_register(struct super_block *sb)
+{
+	struct super_dev *fsd = sb->s_super_dev;
+	int err;
+
+	lockdep_assert_held(&sb_lock);
+	VFS_WARN_ON_ONCE(!sb->s_dev);
+	VFS_WARN_ON_ONCE(!fsd || fsd->sd_dev);
+
+	fsd->sd_dev = sb->s_dev;
+	err = super_dev_insert(fsd);
+	if (err)
+		fsd->sd_dev = 0;
+	return err;
+}
+
+static struct super_dev *super_dev_get(struct rhlist_head *pos)
+{
+	struct super_dev *sb_dev;
+
+	for (; pos; pos = rcu_dereference_all(pos->next)) {
+		sb_dev = container_of(pos, struct super_dev, sd_node);
+		if (refcount_inc_not_zero(&sb_dev->sd_ref))
+			return sb_dev;
+	}
+	return NULL;
+}
+
+static struct super_dev *super_dev_first(dev_t dev)
+{
+	struct super_dev *sb_dev;
+
+	rcu_read_lock();
+	sb_dev = super_dev_get(rhltable_lookup(&super_dev_table, &dev, super_dev_params));
+	rcu_read_unlock();
+	return sb_dev;
+}
+
+static struct super_dev *super_dev_next(struct super_dev *prev)
+{
+	struct super_dev *sb_dev;
+
+	rcu_read_lock();
+	sb_dev = super_dev_get(rcu_dereference_all(prev->sd_node.next));
+	rcu_read_unlock();
+
+	super_dev_put(prev);
+	return sb_dev;
 }
 
 static void kill_super_notify(struct super_block *sb)
@@ -464,6 +569,12 @@ static void kill_super_notify(struct super_block *sb)
 	hlist_del_init(&sb->s_instances);
 	spin_unlock(&sb_lock);
 
+	/* Drop sget_fc()'s claim; a never-registered entry stays with the sb. */
+	if (sb->s_super_dev->sd_dev) {
+		super_dev_put(sb->s_super_dev);
+		sb->s_super_dev = NULL;
+	}
+
 	/*
 	 * Let concurrent mounts know that this thing is really dead.
 	 * We don't need @sb->s_umount here as every concurrent caller
@@ -493,11 +604,7 @@ void deactivate_locked_super(struct super_block *s)
 
 		kill_super_notify(s);
 
-		/*
-		 * Since list_lru_destroy() may sleep, we cannot call it from
-		 * put_super(), where we hold the sb_lock. Therefore we destroy
-		 * the lru lists right now.
-		 */
+		/* list_lru_destroy() may sleep; put_super() callers may not. */
 		list_lru_destroy(&s->s_dentry_lru);
 		list_lru_destroy(&s->s_inode_lru);
 
@@ -544,7 +651,7 @@ static bool grab_super(struct super_block *sb)
 {
 	bool locked;
 
-	sb->s_count++;
+	refcount_inc(&sb->s_passive);
 	spin_unlock(&sb_lock);
 	locked = super_lock_excl(sb);
 	if (locked) {
@@ -571,7 +678,7 @@ static bool grab_super(struct super_block *sb)
  *	lock held in read mode in case of success. On successful return,
  *	the caller must drop the s_umount lock when done.
  *
- *	Note that unlike get_super() et.al. this one does *not* bump ->s_count.
+ *	Note that unlike get_super() et.al. this one does *not* bump ->s_passive.
  *	The reason why it's safe is that we are OK with doing trylock instead
  *	of down_read().  There's a couple of places that are OK with that, but
  *	it's very much not a general-purpose interface.
@@ -778,6 +885,7 @@ retry:
 	}
 	if (!s) {
 		spin_unlock(&sb_lock);
+
 		s = alloc_super(fc->fs_type, fc->sb_flags, user_ns);
 		if (!s)
 			return ERR_PTR(-ENOMEM);
@@ -787,11 +895,13 @@ retry:
 	s->s_fs_info = fc->s_fs_info;
 	err = set(s, fc);
 	if (err) {
+		VFS_WARN_ON_ONCE(s->s_super_dev->sd_dev);
 		s->s_fs_info = NULL;
 		spin_unlock(&sb_lock);
 		destroy_unused_super(s);
 		return ERR_PTR(err);
 	}
+	VFS_WARN_ON_ONCE(!s->s_super_dev->sd_dev);
 	fc->s_fs_info = NULL;
 	s->s_type = fc->fs_type;
 	s->s_iflags |= fc->s_iflags;
@@ -866,14 +976,17 @@ static void __iterate_supers(void (*f)(struct super_block *, void *), void *arg,
 	struct super_block *sb, *p = NULL;
 	bool excl = flags & SUPER_ITER_EXCL;
 
-	guard(spinlock)(&sb_lock);
+	spin_lock(&sb_lock);
 
 	for (sb = first_super(flags);
 	     !list_entry_is_head(sb, &super_blocks, s_list);
 	     sb = next_super(sb, flags)) {
 		if (super_flags(sb, SB_DYING))
 			continue;
-		sb->s_count++;
+
+		if (!refcount_inc_not_zero(&sb->s_passive))
+			continue;
+
 		spin_unlock(&sb_lock);
 
 		if (flags & SUPER_ITER_UNLOCKED) {
@@ -883,13 +996,14 @@ static void __iterate_supers(void (*f)(struct super_block *, void *), void *arg,
 			super_unlock(sb, excl);
 		}
 
-		spin_lock(&sb_lock);
 		if (p)
-			__put_super(p);
+			put_super(p);
 		p = sb;
+		spin_lock(&sb_lock);
 	}
+	spin_unlock(&sb_lock);
 	if (p)
-		__put_super(p);
+		put_super(p);
 }
 
 void iterate_supers(void (*f)(struct super_block *, void *), void *arg)
@@ -918,7 +1032,9 @@ void iterate_supers_type(struct file_system_type *type,
 		if (super_flags(sb, SB_DYING))
 			continue;
 
-		sb->s_count++;
+		if (!refcount_inc_not_zero(&sb->s_passive))
+			continue;
+
 		spin_unlock(&sb_lock);
 
 		locked = super_lock_shared(sb);
@@ -927,41 +1043,33 @@ void iterate_supers_type(struct file_system_type *type,
 			super_unlock_shared(sb);
 		}
 
-		spin_lock(&sb_lock);
 		if (p)
-			__put_super(p);
+			put_super(p);
 		p = sb;
+		spin_lock(&sb_lock);
 	}
-	if (p)
-		__put_super(p);
 	spin_unlock(&sb_lock);
+	if (p)
+		put_super(p);
 }
 
 EXPORT_SYMBOL(iterate_supers_type);
 
 struct super_block *user_get_super(dev_t dev, bool excl)
 {
-	struct super_block *sb;
+	struct super_dev *sb_dev;
 
-	spin_lock(&sb_lock);
-	list_for_each_entry(sb, &super_blocks, s_list) {
-		bool locked;
+	for (sb_dev = super_dev_first(dev); sb_dev; sb_dev = super_dev_next(sb_dev)) {
+		struct super_block *sb = sb_dev->sd_sb;
 
-		if (sb->s_dev != dev)
+		if (!super_lock(sb, excl))
 			continue;
 
-		sb->s_count++;
-		spin_unlock(&sb_lock);
-
-		locked = super_lock(sb, excl);
-		if (locked)
-			return sb;
-
-		spin_lock(&sb_lock);
-		__put_super(sb);
-		break;
+		/* The pinned entry holds a passive reference, take our own. */
+		refcount_inc(&sb->s_passive);
+		super_dev_put(sb_dev);
+		return sb;
 	}
-	spin_unlock(&sb_lock);
 	return NULL;
 }
 
@@ -1243,7 +1351,16 @@ EXPORT_SYMBOL(free_anon_bdev);
 
 int set_anon_super(struct super_block *s, void *data)
 {
-	return get_anon_bdev(&s->s_dev);
+	int error;
+
+	error = get_anon_bdev(&s->s_dev);
+	if (error)
+		return error;
+
+	error = super_dev_register(s);
+	if (error)
+		free_anon_bdev(s->s_dev);
+	return error;
 }
 EXPORT_SYMBOL(set_anon_super);
 
@@ -1329,7 +1446,7 @@ EXPORT_SYMBOL(get_tree_keyed);
 static int set_bdev_super(struct super_block *s, void *data)
 {
 	s->s_dev = *(dev_t *)data;
-	return 0;
+	return super_dev_register(s);
 }
 
 static int super_s_dev_set(struct super_block *s, struct fs_context *fc)
@@ -1371,197 +1488,313 @@ struct super_block *sget_dev(struct fs_context *fc, dev_t dev)
 EXPORT_SYMBOL(sget_dev);
 
 #ifdef CONFIG_BLOCK
-/*
- * Lock the superblock that is holder of the bdev. Returns the superblock
- * pointer if we successfully locked the superblock and it is alive. Otherwise
- * we return NULL and just unlock bdev->bd_holder_lock.
- *
- * The function must be called with bdev->bd_holder_lock and releases it.
- */
-static struct super_block *bdev_super_lock(struct block_device *bdev, bool excl)
-	__releases(&bdev->bd_holder_lock)
+static int fs_super_freeze(struct super_block *sb)
 {
-	struct super_block *sb = bdev->bd_holder;
-	bool locked;
-
-	lockdep_assert_held(&bdev->bd_holder_lock);
-	lockdep_assert_not_held(&sb->s_umount);
-	lockdep_assert_not_held(&bdev->bd_disk->open_mutex);
-
-	/* Make sure sb doesn't go away from under us */
-	spin_lock(&sb_lock);
-	sb->s_count++;
-	spin_unlock(&sb_lock);
-
-	mutex_unlock(&bdev->bd_holder_lock);
-
-	locked = super_lock(sb, excl);
-
-	/*
-	 * If the superblock wasn't already SB_DYING then we hold
-	 * s_umount and can safely drop our temporary reference.
-         */
-	put_super(sb);
-
-	if (!locked)
-		return NULL;
-
-	if (!sb->s_root || !(sb->s_flags & SB_ACTIVE)) {
-		super_unlock(sb, excl);
-		return NULL;
-	}
+	if (sb->s_op->freeze_super)
+		return sb->s_op->freeze_super(sb,
+				FREEZE_MAY_NEST | FREEZE_HOLDER_USERSPACE, NULL);
+	return freeze_super(sb, FREEZE_MAY_NEST | FREEZE_HOLDER_USERSPACE, NULL);
+}
 
-	return sb;
+static int fs_super_thaw(struct super_block *sb)
+{
+	if (sb->s_op->thaw_super)
+		return sb->s_op->thaw_super(sb,
+				FREEZE_MAY_NEST | FREEZE_HOLDER_USERSPACE, NULL);
+	return thaw_super(sb, FREEZE_MAY_NEST | FREEZE_HOLDER_USERSPACE, NULL);
 }
 
 static void fs_bdev_mark_dead(struct block_device *bdev, bool surprise)
 {
-	struct super_block *sb;
+	struct super_dev *sb_dev;
+	dev_t dev = bdev->bd_dev;
 
-	sb = bdev_super_lock(bdev, false);
-	if (!sb)
-		return;
+	mutex_unlock(&bdev->bd_holder_lock);
 
-	if (sb->s_op->remove_bdev) {
-		int ret;
+	for (sb_dev = super_dev_first(dev); sb_dev; sb_dev = super_dev_next(sb_dev)) {
+		struct super_block *sb = sb_dev->sd_sb;
 
-		ret = sb->s_op->remove_bdev(sb, bdev);
-		if (!ret) {
-			super_unlock_shared(sb);
-			return;
+		if (!super_lock_shared(sb))
+			continue;
+		if (sb->s_root && (sb->s_flags & SB_ACTIVE)) {
+			if (!sb->s_op->remove_bdev ||
+			    sb->s_op->remove_bdev(sb, bdev)) {
+				if (!surprise)
+					sync_filesystem(sb);
+				shrink_dcache_sb(sb);
+				evict_inodes(sb);
+				if (sb->s_op->shutdown)
+					sb->s_op->shutdown(sb);
+			}
 		}
-		/* Fallback to shutdown. */
+		super_unlock_shared(sb);
 	}
-
-	if (!surprise)
-		sync_filesystem(sb);
-	shrink_dcache_sb(sb);
-	evict_inodes(sb);
-	if (sb->s_op->shutdown)
-		sb->s_op->shutdown(sb);
-
-	super_unlock_shared(sb);
 }
 
 static void fs_bdev_sync(struct block_device *bdev)
 {
-	struct super_block *sb;
-
-	sb = bdev_super_lock(bdev, false);
-	if (!sb)
-		return;
+	struct super_dev *sb_dev;
+	dev_t dev = bdev->bd_dev;
 
-	sync_filesystem(sb);
-	super_unlock_shared(sb);
-}
+	mutex_unlock(&bdev->bd_holder_lock);
 
-static struct super_block *get_bdev_super(struct block_device *bdev)
-{
-	bool active = false;
-	struct super_block *sb;
+	for (sb_dev = super_dev_first(dev); sb_dev; sb_dev = super_dev_next(sb_dev)) {
+		struct super_block *sb = sb_dev->sd_sb;
 
-	sb = bdev_super_lock(bdev, true);
-	if (sb) {
-		active = atomic_inc_not_zero(&sb->s_active);
-		super_unlock_excl(sb);
+		if (!super_lock_shared(sb))
+			continue;
+		if (sb->s_root && (sb->s_flags & SB_ACTIVE))
+			sync_filesystem(sb);
+		super_unlock_shared(sb);
 	}
-	if (!active)
-		return NULL;
-	return sb;
 }
 
 /**
- * fs_bdev_freeze - freeze owning filesystem of block device
+ * fs_bdev_freeze - freeze every superblock using a block device
  * @bdev: block device
  *
- * Freeze the filesystem that owns this block device if it is still
- * active.
+ * Freeze each live superblock using @bdev.  A superblock owning several block
+ * devices is frozen once per device and stays frozen until all are thawed; the
+ * block layer nests these freezes so the count stays balanced.
  *
- * A filesystem that owns multiple block devices may be frozen from each
- * block device and won't be unfrozen until all block devices are
- * unfrozen. Each block device can only freeze the filesystem once as we
- * nest freezes for block devices in the block layer.
- *
- * Return: If the freeze was successful zero is returned. If the freeze
- *         failed a negative error code is returned.
+ * Return: 0, or the error from the one superblock on a single-fs device.  When
+ *         several superblocks share @bdev a per-superblock failure is swallowed
+ *         (see below), but a sync_blockdev() failure is always reported.
  */
 static int fs_bdev_freeze(struct block_device *bdev)
 {
-	struct super_block *sb;
-	int error = 0;
+	dev_t dev = bdev->bd_dev;
+	struct super_dev *sb_dev;
+	unsigned int count = 0;
+	int error = 0, err;
 
 	lockdep_assert_held(&bdev->bd_fsfreeze_mutex);
 
-	sb = get_bdev_super(bdev);
-	if (!sb)
-		return -EINVAL;
+	mutex_unlock(&bdev->bd_holder_lock);
 
-	if (sb->s_op->freeze_super)
-		error = sb->s_op->freeze_super(sb,
-				FREEZE_MAY_NEST | FREEZE_HOLDER_USERSPACE, NULL);
-	else
-		error = freeze_super(sb,
-				FREEZE_MAY_NEST | FREEZE_HOLDER_USERSPACE, NULL);
+	for (sb_dev = super_dev_first(dev); sb_dev; sb_dev = super_dev_next(sb_dev)) {
+		if (!get_active_super(sb_dev->sd_sb))
+			continue;
+		err = fs_super_freeze(sb_dev->sd_sb);
+		if (err && !error)
+			error = err;
+		deactivate_super(sb_dev->sd_sb);
+		count++;
+	}
+
+	/*
+	 * When several superblocks share the device, keep it frozen even if some
+	 * of them failed to freeze and swallow the error: rolling the rest back
+	 * via thaw_super() can fail too, so neither is a clear win. A single
+	 * filesystem (count == 1) still reports its error.
+	 */
+	if (error && count > 1)
+		error = 0;
 	if (!error)
 		error = sync_blockdev(bdev);
-	deactivate_super(sb);
 	return error;
 }
 
 /**
- * fs_bdev_thaw - thaw owning filesystem of block device
+ * fs_bdev_thaw - thaw every superblock using a block device
  * @bdev: block device
  *
- * Thaw the filesystem that owns this block device.
+ * The counterpart to fs_bdev_freeze(): thaw each live superblock using @bdev.
+ * A zero return does not imply a superblock is fully unfrozen; it may have been
+ * frozen more than once (by the kernel or via another device).
  *
- * A filesystem that owns multiple block devices may be frozen from each
- * block device and won't be unfrozen until all block devices are
- * unfrozen. Each block device can only freeze the filesystem once as we
- * nest freezes for block devices in the block layer.
- *
- * Return: If the thaw was successful zero is returned. If the thaw
- *         failed a negative error code is returned. If this function
- *         returns zero it doesn't mean that the filesystem is unfrozen
- *         as it may have been frozen multiple times (kernel may hold a
- *         freeze or might be frozen from other block devices).
+ * Return: 0, or the first error on a single-fs device; a shared device swallows
+ *         per-superblock errors, as fs_bdev_freeze() does.
  */
 static int fs_bdev_thaw(struct block_device *bdev)
 {
-	struct super_block *sb;
-	int error;
+	dev_t dev = bdev->bd_dev;
+	struct super_dev *sb_dev;
+	unsigned int count = 0;
+	int error = 0, err;
 
 	lockdep_assert_held(&bdev->bd_fsfreeze_mutex);
 
-	/*
-	 * The block device may have been frozen before it was claimed by a
-	 * filesystem. Concurrently another process might try to mount that
-	 * frozen block device and has temporarily claimed the block device for
-	 * that purpose causing a concurrent fs_bdev_thaw() to end up here. The
-	 * mounter is already about to abort mounting because they still saw an
-	 * elevanted bdev->bd_fsfreeze_count so get_bdev_super() will return
-	 * NULL in that case.
-	 */
-	sb = get_bdev_super(bdev);
-	if (!sb)
-		return -EINVAL;
+	mutex_unlock(&bdev->bd_holder_lock);
 
-	if (sb->s_op->thaw_super)
-		error = sb->s_op->thaw_super(sb,
-				FREEZE_MAY_NEST | FREEZE_HOLDER_USERSPACE, NULL);
-	else
-		error = thaw_super(sb,
-				FREEZE_MAY_NEST | FREEZE_HOLDER_USERSPACE, NULL);
-	deactivate_super(sb);
+	for (sb_dev = super_dev_first(dev); sb_dev; sb_dev = super_dev_next(sb_dev)) {
+		if (!get_active_super(sb_dev->sd_sb))
+			continue;
+		err = fs_super_thaw(sb_dev->sd_sb);
+		if (err && !error)
+			error = err;
+		deactivate_super(sb_dev->sd_sb);
+		count++;
+	}
+
+	/* Shared device: swallow per-superblock errors, like fs_bdev_freeze(). */
+	if (error && count > 1)
+		error = 0;
 	return error;
 }
 
-const struct blk_holder_ops fs_holder_ops = {
+static const struct blk_holder_ops fs_holder_ops = {
 	.mark_dead		= fs_bdev_mark_dead,
 	.sync			= fs_bdev_sync,
 	.freeze			= fs_bdev_freeze,
 	.thaw			= fs_bdev_thaw,
 };
-EXPORT_SYMBOL_GPL(fs_holder_ops);
+
+static struct super_dev *super_dev_lookup(dev_t dev, struct super_block *sb)
+{
+	struct super_dev *it;
+	struct rhlist_head *list, *pos;
+
+	RCU_LOCKDEP_WARN(!rcu_read_lock_held(), "suspicious super_dev_lookup() usage");
+	VFS_WARN_ON_ONCE(!dev);
+	VFS_WARN_ON_ONCE(!sb);
+
+	list = rhltable_lookup(&super_dev_table, &dev, super_dev_params);
+	rhl_for_each_entry_rcu(it, pos, list, sd_node) {
+		if (it->sd_sb == sb)
+			return it;
+	}
+
+	return NULL;
+}
+
+static int fs_bdev_register(struct file *bdev_file, struct super_block *sb)
+{
+	struct super_dev *sb_dev __free(kfree) = NULL;
+	dev_t dev = file_bdev(bdev_file)->bd_dev;
+	int err;
+
+	scoped_guard(rcu) {
+		sb_dev = super_dev_lookup(dev, sb);
+		if (sb_dev && refcount_inc_not_zero(&sb_dev->sd_ref)) {
+			retain_and_null_ptr(sb_dev);
+			return 0;
+		}
+	}
+
+	sb_dev = super_dev_alloc(dev, sb);
+	if (!sb_dev)
+		return -ENOMEM;
+
+	err = super_dev_insert(sb_dev);
+	if (err)
+		return err;
+
+	/* Publish the entry before reading the count; pairs with bdev_freeze(). */
+	smp_mb();
+	if (atomic_read(&file_bdev(bdev_file)->bd_fsfreeze_count) > 0) {
+		err = -EBUSY;
+		super_dev_put(sb_dev);
+	}
+
+	retain_and_null_ptr(sb_dev);
+	return err;
+}
+
+/**
+ * fs_bdev_file_open_by_dev - claim a block device on behalf of a superblock
+ * @dev: block device number
+ * @mode: open mode
+ * @holder: block-layer exclusivity token (a superblock, or the file_system_type
+ *          when the device may be shared by several superblocks of that type)
+ * @sb: superblock to drive fs_holder_ops events for
+ *
+ * Open @dev with &fs_holder_ops and register that @sb uses it, so device
+ * removal/sync/freeze/thaw are propagated to @sb (and any other superblock
+ * sharing @dev).  Must be paired with fs_bdev_file_release().
+ *
+ * Return: an opened block-device file or an ERR_PTR().
+ */
+struct file *fs_bdev_file_open_by_dev(dev_t dev, blk_mode_t mode, void *holder,
+				      struct super_block *sb)
+{
+	struct file *bdev_file;
+	int err;
+
+	bdev_file = bdev_file_open_by_dev(dev, mode, holder, &fs_holder_ops);
+	if (IS_ERR(bdev_file))
+		return bdev_file;
+
+	err = fs_bdev_register(bdev_file, sb);
+	if (err) {
+		bdev_fput(bdev_file);
+		return ERR_PTR(err);
+	}
+	return bdev_file;
+}
+EXPORT_SYMBOL_GPL(fs_bdev_file_open_by_dev);
+
+/**
+ * fs_bdev_file_open_by_path - claim a block device on behalf of a superblock
+ * @path: path to the block device
+ * @mode: open mode
+ * @holder: block-layer exclusivity token (a superblock, or the file_system_type
+ *          when the device may be shared by several superblocks of that type)
+ * @sb: superblock to drive fs_holder_ops events for
+ *
+ * Open the block device at @path with &fs_holder_ops and register that @sb
+ * uses it, so device removal/sync/freeze/thaw are propagated to @sb (and any
+ * other superblock sharing the device).  Must be paired with
+ * fs_bdev_file_release().
+ *
+ * Return: an opened block-device file or an ERR_PTR().
+ */
+struct file *fs_bdev_file_open_by_path(const char *path, blk_mode_t mode,
+				       void *holder, struct super_block *sb)
+{
+	struct file *bdev_file;
+	int err;
+
+	bdev_file = bdev_file_open_by_path(path, mode, holder, &fs_holder_ops);
+	if (IS_ERR(bdev_file))
+		return bdev_file;
+
+	err = fs_bdev_register(bdev_file, sb);
+	if (err) {
+		bdev_fput(bdev_file);
+		return ERR_PTR(err);
+	}
+	return bdev_file;
+}
+EXPORT_SYMBOL_GPL(fs_bdev_file_open_by_path);
+
+/**
+ * fs_bdev_unregister - drop a superblock's claim on a block device
+ * @bdev_file: file returned by fs_bdev_file_open_by_{dev,path}()
+ * @sb: superblock the device was claimed for
+ *
+ * The inverse of fs_bdev_register(): drop one claim on the {dev, @sb} entry
+ * (the last claim unregisters it; a pinning cursor defers the actual unlink)
+ * without closing the device.  A caller that must act on the still-open device
+ * between unregistering and closing - e.g. re-allow freezing one denied for a
+ * membership change - pairs this with bdev_fput().  fs_bdev_file_release() is
+ * the common unregister-and-close.
+ */
+void fs_bdev_unregister(struct file *bdev_file, struct super_block *sb)
+{
+	dev_t dev = file_bdev(bdev_file)->bd_dev;
+	struct super_dev *sb_dev;
+
+	rcu_read_lock();
+	sb_dev = super_dev_lookup(dev, sb);
+	rcu_read_unlock();
+	super_dev_put(sb_dev);
+}
+EXPORT_SYMBOL_GPL(fs_bdev_unregister);
+
+/**
+ * fs_bdev_file_release - release a block device claimed for a superblock
+ * @bdev_file: file returned by fs_bdev_file_open_by_{dev,path}()
+ * @sb: superblock the device was claimed for
+ *
+ * Unregister the {dev, @sb} entry, then close the block device.
+ */
+void fs_bdev_file_release(struct file *bdev_file, struct super_block *sb)
+{
+	fs_bdev_unregister(bdev_file, sb);
+	bdev_fput(bdev_file);
+}
+EXPORT_SYMBOL_GPL(fs_bdev_file_release);
 
 int setup_bdev_super(struct super_block *sb, int sb_flags,
 		struct fs_context *fc)
@@ -1570,7 +1803,7 @@ int setup_bdev_super(struct super_block *sb, int sb_flags,
 	struct file *bdev_file;
 	struct block_device *bdev;
 
-	bdev_file = bdev_file_open_by_dev(sb->s_dev, mode, sb, &fs_holder_ops);
+	bdev_file = fs_bdev_file_open_by_dev(sb->s_dev, mode, sb, sb);
 	if (IS_ERR(bdev_file)) {
 		if (fc)
 			errorf(fc, "%s: Can't open blockdev", fc->source);
@@ -1584,20 +1817,19 @@ int setup_bdev_super(struct super_block *sb, int sb_flags,
 	 * writable from userspace even for a read-only block device.
 	 */
 	if ((mode & BLK_OPEN_WRITE) && bdev_read_only(bdev)) {
-		bdev_fput(bdev_file);
+		fs_bdev_file_release(bdev_file, sb);
 		return -EACCES;
 	}
 
-	/*
-	 * It is enough to check bdev was not frozen before we set
-	 * s_bdev as freezing will wait until SB_BORN is set.
-	 */
+	/* The sget_fc() entry is already published; pairs with bdev_freeze(). */
+	smp_mb();
 	if (atomic_read(&bdev->bd_fsfreeze_count) > 0) {
 		if (fc)
 			warnf(fc, "%pg: Can't mount, blockdev is frozen", bdev);
-		bdev_fput(bdev_file);
+		fs_bdev_file_release(bdev_file, sb);
 		return -EBUSY;
 	}
+
 	spin_lock(&sb_lock);
 	sb->s_bdev_file = bdev_file;
 	sb->s_bdev = bdev;
@@ -1686,7 +1918,7 @@ void kill_block_super(struct super_block *sb)
 	generic_shutdown_super(sb);
 	if (bdev) {
 		sync_blockdev(bdev);
-		bdev_fput(sb->s_bdev_file);
+		fs_bdev_file_release(sb->s_bdev_file, sb);
 	}
 }
 
@@ -2033,7 +2265,7 @@ int freeze_super(struct super_block *sb, enum freeze_holder who, const void *fre
 	int ret;
 
 	if (!super_lock_excl(sb)) {
-		WARN_ON_ONCE("Dying superblock while freezing!");
+		WARN_ONCE(1, "Dying superblock while freezing!");
 		return -EINVAL;
 	}
 	atomic_inc(&sb->s_active);
@@ -2197,7 +2429,7 @@ int thaw_super(struct super_block *sb, enum freeze_holder who,
 	       const void *freeze_owner)
 {
 	if (!super_lock_excl(sb)) {
-		WARN_ON_ONCE("Dying superblock while thawing!");
+		WARN_ONCE(1, "Dying superblock while thawing!");
 		return -EINVAL;
 	}
 	return thaw_super_locked(sb, who, freeze_owner);
diff --git a/fs/xfs/xfs_buf.c b/fs/xfs/xfs_buf.c
index 7b1a54da615a..17b9d643e1a8 100644
--- a/fs/xfs/xfs_buf.c
+++ b/fs/xfs/xfs_buf.c
@@ -1631,7 +1631,7 @@ xfs_free_buftarg(
 	fs_put_dax(btp->bt_daxdev, btp->bt_mount);
 	/* the main block device is closed by kill_block_super */
 	if (btp->bt_bdev != btp->bt_mount->m_super->s_bdev)
-		bdev_fput(btp->bt_file);
+		fs_bdev_file_release(btp->bt_file, btp->bt_mount->m_super);
 	kfree(btp);
 }
 
diff --git a/fs/xfs/xfs_super.c b/fs/xfs/xfs_super.c
index 63c4bcbe6c2b..4b2eeb7783f7 100644
--- a/fs/xfs/xfs_super.c
+++ b/fs/xfs/xfs_super.c
@@ -400,8 +400,8 @@ xfs_blkdev_get(
 	blk_mode_t		mode;
 
 	mode = sb_open_mode(mp->m_super->s_flags);
-	*bdev_filep = bdev_file_open_by_path(name, mode,
-			mp->m_super, &fs_holder_ops);
+	*bdev_filep = fs_bdev_file_open_by_path(name, mode,
+			mp->m_super, mp->m_super);
 	if (IS_ERR(*bdev_filep)) {
 		error = PTR_ERR(*bdev_filep);
 		*bdev_filep = NULL;
@@ -526,7 +526,7 @@ xfs_open_devices(
 		mp->m_logdev_targp = mp->m_ddev_targp;
 		/* Handle won't be used, drop it */
 		if (logdev_file)
-			bdev_fput(logdev_file);
+			fs_bdev_file_release(logdev_file, mp->m_super);
 	}
 
 	return 0;
@@ -541,10 +541,10 @@ xfs_open_devices(
 	mp->m_ddev_targp = NULL;
  out_close_rtdev:
 	 if (rtdev_file)
-		bdev_fput(rtdev_file);
+		fs_bdev_file_release(rtdev_file, mp->m_super);
  out_close_logdev:
 	if (logdev_file)
-		bdev_fput(logdev_file);
+		fs_bdev_file_release(logdev_file, mp->m_super);
 	return error;
 }
 
diff --git a/include/linux/blk_types.h b/include/linux/blk_types.h
index 8808ee76e73c..5a725a0cd35f 100644
--- a/include/linux/blk_types.h
+++ b/include/linux/blk_types.h
@@ -66,7 +66,7 @@ struct block_device {
 	int			bd_holders;
 	struct kobject		*bd_holder_dir;
 
-	atomic_t		bd_fsfreeze_count; /* number of freeze requests */
+	atomic_t		bd_fsfreeze_count; /* >0 freeze requests, <0 freeze deniers */
 	struct mutex		bd_fsfreeze_mutex; /* serialize freeze/thaw */
 
 	struct partition_meta_info *bd_meta_info;
diff --git a/include/linux/blkdev.h b/include/linux/blkdev.h
index 9213a5716f95..dbb549cdfb77 100644
--- a/include/linux/blkdev.h
+++ b/include/linux/blkdev.h
@@ -126,8 +126,6 @@ struct blk_integrity {
 	unsigned char				pi_tuple_size;
 };
 
-typedef unsigned int __bitwise blk_mode_t;
-
 /* open for reading */
 #define BLK_OPEN_READ		((__force blk_mode_t)(1 << 0))
 /* open for writing */
@@ -1771,13 +1769,6 @@ struct blk_holder_ops {
 };
 
 /*
- * For filesystems using @fs_holder_ops, the @holder argument passed to
- * helpers used to open and claim block devices via
- * bd_prepare_to_claim() must point to a superblock.
- */
-extern const struct blk_holder_ops fs_holder_ops;
-
-/*
  * Return the correct open flags for blkdev_get_by_* for super block flags
  * as stored in sb->s_flags.
  */
@@ -1837,7 +1828,10 @@ static inline int early_lookup_bdev(const char *pathname, dev_t *dev)
 
 int bdev_freeze(struct block_device *bdev);
 int bdev_thaw(struct block_device *bdev);
+int bdev_deny_freeze(struct block_device *bdev);
+void bdev_allow_freeze(struct block_device *bdev);
 void bdev_fput(struct file *bdev_file);
+void bdev_yield_claim(struct file *bdev_file);
 
 struct io_comp_batch {
 	struct rq_list req_list;
diff --git a/include/linux/fs.h b/include/linux/fs.h
index 2f243b1554d8..04b00f53adc4 100644
--- a/include/linux/fs.h
+++ b/include/linux/fs.h
@@ -1916,8 +1916,6 @@ struct dir_context {
 struct io_uring_cmd;
 struct offset_ctx;
 
-typedef unsigned int __bitwise fop_flags_t;
-
 struct file_operations {
 	struct module *owner;
 	fop_flags_t fop_flags;
diff --git a/include/linux/fs/super.h b/include/linux/fs/super.h
index 405612678115..733d439f01ed 100644
--- a/include/linux/fs/super.h
+++ b/include/linux/fs/super.h
@@ -237,4 +237,12 @@ int thaw_super(struct super_block *super, enum freeze_holder who,
 
 int sb_init_dio_done_wq(struct super_block *sb);
 
+struct file;
+struct file *fs_bdev_file_open_by_dev(dev_t dev, blk_mode_t mode, void *holder,
+				      struct super_block *sb);
+struct file *fs_bdev_file_open_by_path(const char *path, blk_mode_t mode,
+				       void *holder, struct super_block *sb);
+void fs_bdev_unregister(struct file *bdev_file, struct super_block *sb);
+void fs_bdev_file_release(struct file *bdev_file, struct super_block *sb);
+
 #endif /* _LINUX_FS_SUPER_H */
diff --git a/include/linux/fs/super_types.h b/include/linux/fs/super_types.h
index ef7941e9dc79..c8172558750f 100644
--- a/include/linux/fs/super_types.h
+++ b/include/linux/fs/super_types.h
@@ -30,6 +30,7 @@ struct mount;
 struct mtd_info;
 struct quotactl_ops;
 struct shrinker;
+struct super_dev;
 struct unicode_map;
 struct user_namespace;
 struct workqueue_struct;
@@ -132,6 +133,7 @@ struct super_operations {
 struct super_block {
 	struct list_head			s_list;		/* Keep this first */
 	dev_t					s_dev;		/* search index; _not_ kdev_t */
+	struct super_dev			*s_super_dev;	/* sget_fc()'s device table claim */
 	unsigned char				s_blocksize_bits;
 	unsigned long				s_blocksize;
 	loff_t					s_maxbytes;	/* Max file size */
@@ -145,7 +147,7 @@ struct super_block {
 	unsigned long				s_magic;
 	struct dentry				*s_root;
 	struct rw_semaphore			s_umount;
-	int					s_count;
+	refcount_t				s_passive;
 	atomic_t				s_active;
 #ifdef CONFIG_SECURITY
 	void					*s_security;
diff --git a/include/linux/types.h b/include/linux/types.h
index 93166b0b0617..bc5dda2a3d86 100644
--- a/include/linux/types.h
+++ b/include/linux/types.h
@@ -163,6 +163,8 @@ typedef u32 dma_addr_t;
 typedef unsigned int __bitwise gfp_t;
 typedef unsigned int __bitwise slab_flags_t;
 typedef unsigned int __bitwise fmode_t;
+typedef unsigned int __bitwise blk_mode_t;
+typedef unsigned int __bitwise fop_flags_t;
 
 #ifdef CONFIG_PHYS_ADDR_T_64BIT
 typedef u64 phys_addr_t;
diff --git a/tools/testing/selftests/filesystems/.gitignore b/tools/testing/selftests/filesystems/.gitignore
index a78f894157de..9eb185fb2f9d 100644
--- a/tools/testing/selftests/filesystems/.gitignore
+++ b/tools/testing/selftests/filesystems/.gitignore
@@ -6,3 +6,4 @@ file_stressor
 anon_inode_test
 kernfs_test
 idmapped_tmpfile
+ustat_test
diff --git a/tools/testing/selftests/filesystems/Makefile b/tools/testing/selftests/filesystems/Makefile
index a7ec2ba2dd83..03be337c1f35 100644
--- a/tools/testing/selftests/filesystems/Makefile
+++ b/tools/testing/selftests/filesystems/Makefile
@@ -1,7 +1,7 @@
 # SPDX-License-Identifier: GPL-2.0
 
 CFLAGS += $(KHDR_INCLUDES)
-TEST_GEN_PROGS := devpts_pts file_stressor anon_inode_test kernfs_test fclog
+TEST_GEN_PROGS := devpts_pts file_stressor anon_inode_test kernfs_test fclog ustat_test
 TEST_GEN_PROGS += idmapped_tmpfile
 TEST_GEN_PROGS_EXTENDED := dnotify_test
 
diff --git a/tools/testing/selftests/filesystems/ustat_test.c b/tools/testing/selftests/filesystems/ustat_test.c
new file mode 100644
index 000000000000..d429fd18d779
--- /dev/null
+++ b/tools/testing/selftests/filesystems/ustat_test.c
@@ -0,0 +1,135 @@
+// SPDX-License-Identifier: GPL-2.0
+/*
+ * Test ustat(2): looking up superblocks by device number.
+ *
+ * ustat() resolves a device number to a mounted superblock via
+ * user_get_super(). Check that the device number of a mounted tmpfs (an
+ * anonymous device) resolves, that it stops resolving once the filesystem
+ * is unmounted and that bogus device numbers report EINVAL.
+ */
+#define _GNU_SOURCE
+#include <errno.h>
+#include <fcntl.h>
+#include <sched.h>
+#include <stdio.h>
+#include <stdlib.h>
+#include <string.h>
+#include <sys/mount.h>
+#include <sys/stat.h>
+#include <sys/syscall.h>
+#include <unistd.h>
+
+#include "../kselftest_harness.h"
+
+/* struct ustat is not exported through UAPI, mirror include/linux/types.h. */
+struct ustat_buf {
+	int		f_tfree;
+	unsigned long	f_tinode;
+	char		f_fname[6];
+	char		f_fpack[6];
+	/* slack in case an architecture lays the struct out differently */
+	char		pad[64];
+};
+
+#ifdef __NR_ustat
+
+/*
+ * The kernel decodes @dev with new_decode_dev(), which matches the low 32
+ * bits of the st_dev encoding stat(2) returns for any major below 4096.
+ */
+static int sys_ustat(unsigned int dev, struct ustat_buf *buf)
+{
+	return syscall(__NR_ustat, dev, buf);
+}
+
+static int write_string(const char *path, const char *string)
+{
+	ssize_t len = strlen(string);
+	int fd;
+
+	fd = open(path, O_WRONLY);
+	if (fd < 0)
+		return -1;
+	if (write(fd, string, len) != len) {
+		close(fd);
+		return -1;
+	}
+	return close(fd);
+}
+
+/* Enter namespaces in which mounting a tmpfs instance is allowed. */
+static int setup_namespaces(void)
+{
+	uid_t uid = getuid();
+	gid_t gid = getgid();
+	char map[64];
+
+	if (unshare(CLONE_NEWNS | (uid ? CLONE_NEWUSER : 0)))
+		return -1;
+
+	if (uid) {
+		if (write_string("/proc/self/setgroups", "deny"))
+			return -1;
+		snprintf(map, sizeof(map), "0 %d 1", uid);
+		if (write_string("/proc/self/uid_map", map))
+			return -1;
+		snprintf(map, sizeof(map), "0 %d 1", gid);
+		if (write_string("/proc/self/gid_map", map))
+			return -1;
+	}
+
+	return mount(NULL, "/", NULL, MS_REC | MS_PRIVATE, NULL);
+}
+
+TEST(resolves_mounted_superblock)
+{
+	char dir[] = "/tmp/ustat_test.XXXXXX";
+	struct ustat_buf ub;
+	struct stat st;
+
+	ASSERT_NE(NULL, mkdtemp(dir));
+
+	if (setup_namespaces()) {
+		rmdir(dir);
+		SKIP(return, "cannot set up namespaces: %s", strerror(errno));
+	}
+
+	ASSERT_EQ(0, mount("ustat_test", dir, "tmpfs", 0, NULL));
+	ASSERT_EQ(0, stat(dir, &st));
+
+	memset(&ub, 0xff, sizeof(ub));
+	ASSERT_EQ(0, sys_ustat(st.st_dev, &ub))
+		TH_LOG("ustat(%u): %s", (unsigned int)st.st_dev,
+		       strerror(errno));
+
+	ASSERT_EQ(0, umount(dir));
+
+	/* The unmount removed the superblock, the device is gone. */
+	ASSERT_EQ(-1, sys_ustat(st.st_dev, &ub));
+	ASSERT_EQ(EINVAL, errno);
+
+	rmdir(dir);
+}
+
+TEST(bogus_device_numbers)
+{
+	struct ustat_buf ub;
+
+	ASSERT_EQ(-1, sys_ustat(0, &ub));
+	ASSERT_EQ(EINVAL, errno);
+
+	/* major 4095, minor 1048575: nothing plausible lives there */
+	ASSERT_EQ(-1, sys_ustat((0xfffu << 8) | 0xffu | (0xfff00u << 12), &ub));
+	ASSERT_EQ(EINVAL, errno);
+}
+
+#else /* !__NR_ustat */
+
+TEST(unsupported)
+{
+	SKIP(return, "ustat(2) is not available on this architecture");
+}
+
+#endif /* __NR_ustat */
+
+TEST_HARNESS_MAIN