Merge tag 'for-5.11-rc2-tag' of git://git.kernel.org/pub/scm/linux/kernel/git/kdave/linuxHEADmaster

Pull btrfs fixes from David Sterba:
 "A few more fixes that arrived before the end of the year:

   - a bunch of fixes related to transaction handle lifetime wrt various
     operations (umount, remount, qgroup scan, orphan cleanup)

   - async discard scheduling fixes

   - fix item size calculation when item keys collide for extend refs
     (hardlinks)

   - fix qgroup flushing from running transaction

   - fix send, wrong file path when there is an inode with a pending
     rmdir

   - fix deadlock when cloning inline extent and low on free metadata
     space"

* tag 'for-5.11-rc2-tag' of git://git.kernel.org/pub/scm/linux/kernel/git/kdave/linux:
  btrfs: run delayed iputs when remounting RO to avoid leaking them
  btrfs: add assertion for empty list of transactions at late stage of umount
  btrfs: fix race between RO remount and the cleaner task
  btrfs: fix transaction leak and crash after cleaning up orphans on RO mount
  btrfs: fix transaction leak and crash after RO remount caused by qgroup rescan
  btrfs: merge critical sections of discard lock in workfn
  btrfs: fix racy access to discard_ctl data
  btrfs: fix async discard stall
  btrfs: tests: initialize test inodes location
  btrfs: send: fix wrong file path when there is an inode with a pending rmdir
  btrfs: qgroup: don't try to wait flushing if we're already holding a transaction
  btrfs: correctly calculate item size used when item key collision happens
  btrfs: fix deadlock when cloning inline extent and low on free metadata space

18 files changed, 243 insertions, 97 deletions

diff --git a/fs/btrfs/btrfs_inode.h b/fs/btrfs/btrfs_inode.h
index 555cbcef6585..d9bf53d9ff90 100644
--- a/fs/btrfs/btrfs_inode.h
+++ b/fs/btrfs/btrfs_inode.h
@@ -42,6 +42,15 @@ enum {
 	 * to an inode.
 	 */
 	BTRFS_INODE_NO_XATTRS,
+	/*
+	 * Set when we are in a context where we need to start a transaction and
+	 * have dirty pages with the respective file range locked. This is to
+	 * ensure that when reserving space for the transaction, if we are low
+	 * on available space and need to flush delalloc, we will not flush
+	 * delalloc for this inode, because that could result in a deadlock (on
+	 * the file range, inode's io_tree).
+	 */
+	BTRFS_INODE_NO_DELALLOC_FLUSH,
 };
 
 /* in memory btrfs inode */
diff --git a/fs/btrfs/ctree.c b/fs/btrfs/ctree.c
index 07810891e204..cc89b63d65a4 100644
--- a/fs/btrfs/ctree.c
+++ b/fs/btrfs/ctree.c
@@ -2555,8 +2555,14 @@ out:
  * @p:		Holds all btree nodes along the search path
  * @root:	The root node of the tree
  * @key:	The key we are looking for
- * @ins_len:	Indicates purpose of search, for inserts it is 1, for
- *		deletions it's -1. 0 for plain searches
+ * @ins_len:	Indicates purpose of search:
+ *              >0  for inserts it's size of item inserted (*)
+ *              <0  for deletions
+ *               0  for plain searches, not modifying the tree
+ *
+ *              (*) If size of item inserted doesn't include
+ *              sizeof(struct btrfs_item), then p->search_for_extension must
+ *              be set.
  * @cow:	boolean should CoW operations be performed. Must always be 1
  *		when modifying the tree.
  *
@@ -2717,6 +2723,20 @@ cow_done:
 
 		if (level == 0) {
 			p->slots[level] = slot;
+			/*
+			 * Item key already exists. In this case, if we are
+			 * allowed to insert the item (for example, in dir_item
+			 * case, item key collision is allowed), it will be
+			 * merged with the original item. Only the item size
+			 * grows, no new btrfs item will be added. If
+			 * search_for_extension is not set, ins_len already
+			 * accounts the size btrfs_item, deduct it here so leaf
+			 * space check will be correct.
+			 */
+			if (ret == 0 && ins_len > 0 && !p->search_for_extension) {
+				ASSERT(ins_len >= sizeof(struct btrfs_item));
+				ins_len -= sizeof(struct btrfs_item);
+			}
 			if (ins_len > 0 &&
 			    btrfs_leaf_free_space(b) < ins_len) {
 				if (write_lock_level < 1) {
diff --git a/fs/btrfs/ctree.h b/fs/btrfs/ctree.h
index 1d3c1e479f3d..e6e37591f1de 100644
--- a/fs/btrfs/ctree.h
+++ b/fs/btrfs/ctree.h
@@ -131,6 +131,8 @@ enum {
 	 * defrag
 	 */
 	BTRFS_FS_STATE_REMOUNTING,
+	/* Filesystem in RO mode */
+	BTRFS_FS_STATE_RO,
 	/* Track if a transaction abort has been reported on this filesystem */
 	BTRFS_FS_STATE_TRANS_ABORTED,
 	/*
@@ -367,6 +369,12 @@ struct btrfs_path {
 	unsigned int search_commit_root:1;
 	unsigned int need_commit_sem:1;
 	unsigned int skip_release_on_error:1;
+	/*
+	 * Indicate that new item (btrfs_search_slot) is extending already
+	 * existing item and ins_len contains only the data size and not item
+	 * header (ie. sizeof(struct btrfs_item) is not included).
+	 */
+	unsigned int search_for_extension:1;
 };
 #define BTRFS_MAX_EXTENT_ITEM_SIZE(r) ((BTRFS_LEAF_DATA_SIZE(r->fs_info) >> 4) - \
 					sizeof(struct btrfs_item))
@@ -2885,10 +2893,26 @@ static inline int btrfs_fs_closing(struct btrfs_fs_info *fs_info)
  * If we remount the fs to be R/O or umount the fs, the cleaner needn't do
  * anything except sleeping. This function is used to check the status of
  * the fs.
+ * We check for BTRFS_FS_STATE_RO to avoid races with a concurrent remount,
+ * since setting and checking for SB_RDONLY in the superblock's flags is not
+ * atomic.
  */
 static inline int btrfs_need_cleaner_sleep(struct btrfs_fs_info *fs_info)
 {
-	return fs_info->sb->s_flags & SB_RDONLY || btrfs_fs_closing(fs_info);
+	return test_bit(BTRFS_FS_STATE_RO, &fs_info->fs_state) ||
+		btrfs_fs_closing(fs_info);
+}
+
+static inline void btrfs_set_sb_rdonly(struct super_block *sb)
+{
+	sb->s_flags |= SB_RDONLY;
+	set_bit(BTRFS_FS_STATE_RO, &btrfs_sb(sb)->fs_state);
+}
+
+static inline void btrfs_clear_sb_rdonly(struct super_block *sb)
+{
+	sb->s_flags &= ~SB_RDONLY;
+	clear_bit(BTRFS_FS_STATE_RO, &btrfs_sb(sb)->fs_state);
 }
 
 /* tree mod log functions from ctree.c */
@@ -3073,7 +3097,8 @@ int btrfs_truncate_inode_items(struct btrfs_trans_handle *trans,
 			       u32 min_type);
 
 int btrfs_start_delalloc_snapshot(struct btrfs_root *root);
-int btrfs_start_delalloc_roots(struct btrfs_fs_info *fs_info, u64 nr);
+int btrfs_start_delalloc_roots(struct btrfs_fs_info *fs_info, u64 nr,
+			       bool in_reclaim_context);
 int btrfs_set_extent_delalloc(struct btrfs_inode *inode, u64 start, u64 end,
 			      unsigned int extra_bits,
 			      struct extent_state **cached_state);
diff --git a/fs/btrfs/dev-replace.c b/fs/btrfs/dev-replace.c
index a98e33f232d5..324f646d6e5e 100644
--- a/fs/btrfs/dev-replace.c
+++ b/fs/btrfs/dev-replace.c
@@ -715,7 +715,7 @@ static int btrfs_dev_replace_finishing(struct btrfs_fs_info *fs_info,
 	 * flush all outstanding I/O and inode extent mappings before the
 	 * copy operation is declared as being finished
 	 */
-	ret = btrfs_start_delalloc_roots(fs_info, U64_MAX);
+	ret = btrfs_start_delalloc_roots(fs_info, U64_MAX, false);
 	if (ret) {
 		mutex_unlock(&dev_replace->lock_finishing_cancel_unmount);
 		return ret;
diff --git a/fs/btrfs/discard.c b/fs/btrfs/discard.c
index 1db966bf85b2..2b8383d41144 100644
--- a/fs/btrfs/discard.c
+++ b/fs/btrfs/discard.c
@@ -199,16 +199,15 @@ static struct btrfs_block_group *find_next_block_group(
 static struct btrfs_block_group *peek_discard_list(
 					struct btrfs_discard_ctl *discard_ctl,
 					enum btrfs_discard_state *discard_state,
-					int *discard_index)
+					int *discard_index, u64 now)
 {
 	struct btrfs_block_group *block_group;
-	const u64 now = ktime_get_ns();
 
 	spin_lock(&discard_ctl->lock);
 again:
 	block_group = find_next_block_group(discard_ctl, now);
 
-	if (block_group && now > block_group->discard_eligible_time) {
+	if (block_group && now >= block_group->discard_eligible_time) {
 		if (block_group->discard_index == BTRFS_DISCARD_INDEX_UNUSED &&
 		    block_group->used != 0) {
 			if (btrfs_is_block_group_data_only(block_group))
@@ -222,12 +221,11 @@ again:
 			block_group->discard_state = BTRFS_DISCARD_EXTENTS;
 		}
 		discard_ctl->block_group = block_group;
+	}
+	if (block_group) {
 		*discard_state = block_group->discard_state;
 		*discard_index = block_group->discard_index;
-	} else {
-		block_group = NULL;
 	}
-
 	spin_unlock(&discard_ctl->lock);
 
 	return block_group;
@@ -330,28 +328,15 @@ void btrfs_discard_queue_work(struct btrfs_discard_ctl *discard_ctl,
 		btrfs_discard_schedule_work(discard_ctl, false);
 }
 
-/**
- * btrfs_discard_schedule_work - responsible for scheduling the discard work
- * @discard_ctl: discard control
- * @override: override the current timer
- *
- * Discards are issued by a delayed workqueue item.  @override is used to
- * update the current delay as the baseline delay interval is reevaluated on
- * transaction commit.  This is also maxed with any other rate limit.
- */
-void btrfs_discard_schedule_work(struct btrfs_discard_ctl *discard_ctl,
-				 bool override)
+static void __btrfs_discard_schedule_work(struct btrfs_discard_ctl *discard_ctl,
+					  u64 now, bool override)
 {
 	struct btrfs_block_group *block_group;
-	const u64 now = ktime_get_ns();
-
-	spin_lock(&discard_ctl->lock);
 
 	if (!btrfs_run_discard_work(discard_ctl))
-		goto out;
-
+		return;
 	if (!override && delayed_work_pending(&discard_ctl->work))
-		goto out;
+		return;
 
 	block_group = find_next_block_group(discard_ctl, now);
 	if (block_group) {
@@ -393,7 +378,24 @@ void btrfs_discard_schedule_work(struct btrfs_discard_ctl *discard_ctl,
 		mod_delayed_work(discard_ctl->discard_workers,
 				 &discard_ctl->work, nsecs_to_jiffies(delay));
 	}
-out:
+}
+
+/*
+ * btrfs_discard_schedule_work - responsible for scheduling the discard work
+ * @discard_ctl:  discard control
+ * @override:     override the current timer
+ *
+ * Discards are issued by a delayed workqueue item.  @override is used to
+ * update the current delay as the baseline delay interval is reevaluated on
+ * transaction commit.  This is also maxed with any other rate limit.
+ */
+void btrfs_discard_schedule_work(struct btrfs_discard_ctl *discard_ctl,
+				 bool override)
+{
+	const u64 now = ktime_get_ns();
+
+	spin_lock(&discard_ctl->lock);
+	__btrfs_discard_schedule_work(discard_ctl, now, override);
 	spin_unlock(&discard_ctl->lock);
 }
 
@@ -438,13 +440,18 @@ static void btrfs_discard_workfn(struct work_struct *work)
 	int discard_index = 0;
 	u64 trimmed = 0;
 	u64 minlen = 0;
+	u64 now = ktime_get_ns();
 
 	discard_ctl = container_of(work, struct btrfs_discard_ctl, work.work);
 
 	block_group = peek_discard_list(discard_ctl, &discard_state,
-					&discard_index);
+					&discard_index, now);
 	if (!block_group || !btrfs_run_discard_work(discard_ctl))
 		return;
+	if (now < block_group->discard_eligible_time) {
+		btrfs_discard_schedule_work(discard_ctl, false);
+		return;
+	}
 
 	/* Perform discarding */
 	minlen = discard_minlen[discard_index];
@@ -474,13 +481,6 @@ static void btrfs_discard_workfn(struct work_struct *work)
 		discard_ctl->discard_extent_bytes += trimmed;
 	}
 
-	/*
-	 * Updated without locks as this is inside the workfn and nothing else
-	 * is reading the values
-	 */
-	discard_ctl->prev_discard = trimmed;
-	discard_ctl->prev_discard_time = ktime_get_ns();
-
 	/* Determine next steps for a block_group */
 	if (block_group->discard_cursor >= btrfs_block_group_end(block_group)) {
 		if (discard_state == BTRFS_DISCARD_BITMAPS) {
@@ -496,11 +496,13 @@ static void btrfs_discard_workfn(struct work_struct *work)
 		}
 	}
 
+	now = ktime_get_ns();
 	spin_lock(&discard_ctl->lock);
+	discard_ctl->prev_discard = trimmed;
+	discard_ctl->prev_discard_time = now;
 	discard_ctl->block_group = NULL;
+	__btrfs_discard_schedule_work(discard_ctl, now, false);
 	spin_unlock(&discard_ctl->lock);
-
-	btrfs_discard_schedule_work(discard_ctl, false);
 }
 
 /**
diff --git a/fs/btrfs/disk-io.c b/fs/btrfs/disk-io.c
index 765deefda92b..1dfd4b2d0e1e 100644
--- a/fs/btrfs/disk-io.c
+++ b/fs/btrfs/disk-io.c
@@ -1729,7 +1729,7 @@ static int cleaner_kthread(void *arg)
 		 */
 		btrfs_delete_unused_bgs(fs_info);
 sleep:
-		clear_bit(BTRFS_FS_CLEANER_RUNNING, &fs_info->flags);
+		clear_and_wake_up_bit(BTRFS_FS_CLEANER_RUNNING, &fs_info->flags);
 		if (kthread_should_park())
 			kthread_parkme();
 		if (kthread_should_stop())
@@ -2830,6 +2830,9 @@ static int init_mount_fs_info(struct btrfs_fs_info *fs_info, struct super_block
 		return -ENOMEM;
 	btrfs_init_delayed_root(fs_info->delayed_root);
 
+	if (sb_rdonly(sb))
+		set_bit(BTRFS_FS_STATE_RO, &fs_info->fs_state);
+
 	return btrfs_alloc_stripe_hash_table(fs_info);
 }
 
@@ -2969,6 +2972,7 @@ int btrfs_start_pre_rw_mount(struct btrfs_fs_info *fs_info)
 		}
 	}
 
+	ret = btrfs_find_orphan_roots(fs_info);
 out:
 	return ret;
 }
@@ -3383,10 +3387,6 @@ int __cold open_ctree(struct super_block *sb, struct btrfs_fs_devices *fs_device
 		}
 	}
 
-	ret = btrfs_find_orphan_roots(fs_info);
-	if (ret)
-		goto fail_qgroup;
-
 	fs_info->fs_root = btrfs_get_fs_root(fs_info, BTRFS_FS_TREE_OBJECTID, true);
 	if (IS_ERR(fs_info->fs_root)) {
 		err = PTR_ERR(fs_info->fs_root);
@@ -4181,6 +4181,9 @@ void __cold close_ctree(struct btrfs_fs_info *fs_info)
 	invalidate_inode_pages2(fs_info->btree_inode->i_mapping);
 	btrfs_stop_all_workers(fs_info);
 
+	/* We shouldn't have any transaction open at this point */
+	ASSERT(list_empty(&fs_info->trans_list));
+
 	clear_bit(BTRFS_FS_OPEN, &fs_info->flags);
 	free_root_pointers(fs_info, true);
 	btrfs_free_fs_roots(fs_info);
diff --git a/fs/btrfs/extent-tree.c b/fs/btrfs/extent-tree.c
index 56ea380f5a17..d79b8369e6aa 100644
--- a/fs/btrfs/extent-tree.c
+++ b/fs/btrfs/extent-tree.c
@@ -844,6 +844,7 @@ int lookup_inline_extent_backref(struct btrfs_trans_handle *trans,
 	want = extent_ref_type(parent, owner);
 	if (insert) {
 		extra_size = btrfs_extent_inline_ref_size(want);
+		path->search_for_extension = 1;
 		path->keep_locks = 1;
 	} else
 		extra_size = -1;
@@ -996,6 +997,7 @@ again:
 out:
 	if (insert) {
 		path->keep_locks = 0;
+		path->search_for_extension = 0;
 		btrfs_unlock_up_safe(path, 1);
 	}
 	return err;
diff --git a/fs/btrfs/file-item.c b/fs/btrfs/file-item.c
index 1545c22ef280..6ccfc019ad90 100644
--- a/fs/btrfs/file-item.c
+++ b/fs/btrfs/file-item.c
@@ -1016,8 +1016,10 @@ again:
 	}
 
 	btrfs_release_path(path);
+	path->search_for_extension = 1;
 	ret = btrfs_search_slot(trans, root, &file_key, path,
 				csum_size, 1);
+	path->search_for_extension = 0;
 	if (ret < 0)
 		goto out;
 
diff --git a/fs/btrfs/inode.c b/fs/btrfs/inode.c
index 8e23780acfae..070716650df8 100644
--- a/fs/btrfs/inode.c
+++ b/fs/btrfs/inode.c
@@ -9390,7 +9390,8 @@ static struct btrfs_delalloc_work *btrfs_alloc_delalloc_work(struct inode *inode
  * some fairly slow code that needs optimization. This walks the list
  * of all the inodes with pending delalloc and forces them to disk.
  */
-static int start_delalloc_inodes(struct btrfs_root *root, u64 *nr, bool snapshot)
+static int start_delalloc_inodes(struct btrfs_root *root, u64 *nr, bool snapshot,
+				 bool in_reclaim_context)
 {
 	struct btrfs_inode *binode;
 	struct inode *inode;
@@ -9411,6 +9412,11 @@ static int start_delalloc_inodes(struct btrfs_root *root, u64 *nr, bool snapshot
 
 		list_move_tail(&binode->delalloc_inodes,
 			       &root->delalloc_inodes);
+
+		if (in_reclaim_context &&
+		    test_bit(BTRFS_INODE_NO_DELALLOC_FLUSH, &binode->runtime_flags))
+			continue;
+
 		inode = igrab(&binode->vfs_inode);
 		if (!inode) {
 			cond_resched_lock(&root->delalloc_lock);
@@ -9464,10 +9470,11 @@ int btrfs_start_delalloc_snapshot(struct btrfs_root *root)
 	if (test_bit(BTRFS_FS_STATE_ERROR, &fs_info->fs_state))
 		return -EROFS;
 
-	return start_delalloc_inodes(root, &nr, true);
+	return start_delalloc_inodes(root, &nr, true, false);
 }
 
-int btrfs_start_delalloc_roots(struct btrfs_fs_info *fs_info, u64 nr)
+int btrfs_start_delalloc_roots(struct btrfs_fs_info *fs_info, u64 nr,
+			       bool in_reclaim_context)
 {
 	struct btrfs_root *root;
 	struct list_head splice;
@@ -9490,7 +9497,7 @@ int btrfs_start_delalloc_roots(struct btrfs_fs_info *fs_info, u64 nr)
 			       &fs_info->delalloc_roots);
 		spin_unlock(&fs_info->delalloc_root_lock);
 
-		ret = start_delalloc_inodes(root, &nr, false);
+		ret = start_delalloc_inodes(root, &nr, false, in_reclaim_context);
 		btrfs_put_root(root);
 		if (ret < 0)
 			goto out;
diff --git a/fs/btrfs/ioctl.c b/fs/btrfs/ioctl.c
index 703212ff50a5..dde49a791f3e 100644
--- a/fs/btrfs/ioctl.c
+++ b/fs/btrfs/ioctl.c
@@ -4951,7 +4951,7 @@ long btrfs_ioctl(struct file *file, unsigned int
 	case BTRFS_IOC_SYNC: {
 		int ret;
 
-		ret = btrfs_start_delalloc_roots(fs_info, U64_MAX);
+		ret = btrfs_start_delalloc_roots(fs_info, U64_MAX, false);
 		if (ret)
 			return ret;
 		ret = btrfs_sync_fs(inode->i_sb, 1);
diff --git a/fs/btrfs/qgroup.c b/fs/btrfs/qgroup.c
index fe3046007f52..808370ada888 100644
--- a/fs/btrfs/qgroup.c
+++ b/fs/btrfs/qgroup.c
@@ -3190,6 +3190,12 @@ out:
 	return ret;
 }
 
+static bool rescan_should_stop(struct btrfs_fs_info *fs_info)
+{
+	return btrfs_fs_closing(fs_info) ||
+		test_bit(BTRFS_FS_STATE_REMOUNTING, &fs_info->fs_state);
+}
+
 static void btrfs_qgroup_rescan_worker(struct btrfs_work *work)
 {
 	struct btrfs_fs_info *fs_info = container_of(work, struct btrfs_fs_info,
@@ -3198,6 +3204,7 @@ static void btrfs_qgroup_rescan_worker(struct btrfs_work *work)
 	struct btrfs_trans_handle *trans = NULL;
 	int err = -ENOMEM;
 	int ret = 0;
+	bool stopped = false;
 
 	path = btrfs_alloc_path();
 	if (!path)
@@ -3210,7 +3217,7 @@ static void btrfs_qgroup_rescan_worker(struct btrfs_work *work)
 	path->skip_locking = 1;
 
 	err = 0;
-	while (!err && !btrfs_fs_closing(fs_info)) {
+	while (!err && !(stopped = rescan_should_stop(fs_info))) {
 		trans = btrfs_start_transaction(fs_info->fs_root, 0);
 		if (IS_ERR(trans)) {
 			err = PTR_ERR(trans);
@@ -3253,7 +3260,7 @@ out:
 	}
 
 	mutex_lock(&fs_info->qgroup_rescan_lock);
-	if (!btrfs_fs_closing(fs_info))
+	if (!stopped)
 		fs_info->qgroup_flags &= ~BTRFS_QGROUP_STATUS_FLAG_RESCAN;
 	if (trans) {
 		ret = update_qgroup_status_item(trans);
@@ -3272,7 +3279,7 @@ out:
 
 	btrfs_end_transaction(trans);
 
-	if (btrfs_fs_closing(fs_info)) {
+	if (stopped) {
 		btrfs_info(fs_info, "qgroup scan paused");
 	} else if (err >= 0) {
 		btrfs_info(fs_info, "qgroup scan completed%s",
@@ -3531,16 +3538,6 @@ static int try_flush_qgroup(struct btrfs_root *root)
 	bool can_commit = true;
 
 	/*
-	 * We don't want to run flush again and again, so if there is a running
-	 * one, we won't try to start a new flush, but exit directly.
-	 */
-	if (test_and_set_bit(BTRFS_ROOT_QGROUP_FLUSHING, &root->state)) {
-		wait_event(root->qgroup_flush_wait,
-			!test_bit(BTRFS_ROOT_QGROUP_FLUSHING, &root->state));
-		return 0;
-	}
-
-	/*
 	 * If current process holds a transaction, we shouldn't flush, as we
 	 * assume all space reservation happens before a transaction handle is
 	 * held.
@@ -3554,6 +3551,26 @@ static int try_flush_qgroup(struct btrfs_root *root)
 	    current->journal_info != BTRFS_SEND_TRANS_STUB)
 		can_commit = false;
 
+	/*
+	 * We don't want to run flush again and again, so if there is a running
+	 * one, we won't try to start a new flush, but exit directly.
+	 */
+	if (test_and_set_bit(BTRFS_ROOT_QGROUP_FLUSHING, &root->state)) {
+		/*
+		 * We are already holding a transaction, thus we can block other
+		 * threads from flushing.  So exit right now. This increases
+		 * the chance of EDQUOT for heavy load and near limit cases.
+		 * But we can argue that if we're already near limit, EDQUOT is
+		 * unavoidable anyway.
+		 */
+		if (!can_commit)
+			return 0;
+
+		wait_event(root->qgroup_flush_wait,
+			!test_bit(BTRFS_ROOT_QGROUP_FLUSHING, &root->state));
+		return 0;
+	}
+
 	ret = btrfs_start_delalloc_snapshot(root);
 	if (ret < 0)
 		goto out;
diff --git a/fs/btrfs/reflink.c b/fs/btrfs/reflink.c
index ab80896315be..b03e7891394e 100644
--- a/fs/btrfs/reflink.c
+++ b/fs/btrfs/reflink.c
@@ -89,6 +89,19 @@ static int copy_inline_to_page(struct btrfs_inode *inode,
 	if (ret)
 		goto out_unlock;
 
+	/*
+	 * After dirtying the page our caller will need to start a transaction,
+	 * and if we are low on metadata free space, that can cause flushing of
+	 * delalloc for all inodes in order to get metadata space released.
+	 * However we are holding the range locked for the whole duration of
+	 * the clone/dedupe operation, so we may deadlock if that happens and no
+	 * other task releases enough space. So mark this inode as not being
+	 * possible to flush to avoid such deadlock. We will clear that flag
+	 * when we finish cloning all extents, since a transaction is started
+	 * after finding each extent to clone.
+	 */
+	set_bit(BTRFS_INODE_NO_DELALLOC_FLUSH, &inode->runtime_flags);
+
 	if (comp_type == BTRFS_COMPRESS_NONE) {
 		char *map;
 
@@ -549,6 +562,8 @@ process_slot:
 out:
 	btrfs_free_path(path);
 	kvfree(buf);
+	clear_bit(BTRFS_INODE_NO_DELALLOC_FLUSH, &BTRFS_I(inode)->runtime_flags);
+
 	return ret;
 }
 
diff --git a/fs/btrfs/send.c b/fs/btrfs/send.c
index d719a2755a40..ae97f4dbaff3 100644
--- a/fs/btrfs/send.c
+++ b/fs/btrfs/send.c
@@ -236,6 +236,7 @@ struct waiting_dir_move {
 	 * after this directory is moved, we can try to rmdir the ino rmdir_ino.
 	 */
 	u64 rmdir_ino;
+	u64 rmdir_gen;
 	bool orphanized;
 };
 
@@ -316,7 +317,7 @@ static int is_waiting_for_move(struct send_ctx *sctx, u64 ino);
 static struct waiting_dir_move *
 get_waiting_dir_move(struct send_ctx *sctx, u64 ino);
 
-static int is_waiting_for_rm(struct send_ctx *sctx, u64 dir_ino);
+static int is_waiting_for_rm(struct send_ctx *sctx, u64 dir_ino, u64 gen);
 
 static int need_send_hole(struct send_ctx *sctx)
 {
@@ -2299,7 +2300,7 @@ static int get_cur_path(struct send_ctx *sctx, u64 ino, u64 gen,
 
 		fs_path_reset(name);
 
-		if (is_waiting_for_rm(sctx, ino)) {
+		if (is_waiting_for_rm(sctx, ino, gen)) {
 			ret = gen_unique_name(sctx, ino, gen, name);
 			if (ret < 0)
 				goto out;
@@ -2858,8 +2859,8 @@ out:
 	return ret;
 }
 
-static struct orphan_dir_info *
-add_orphan_dir_info(struct send_ctx *sctx, u64 dir_ino)
+static struct orphan_dir_info *add_orphan_dir_info(struct send_ctx *sctx,
+						   u64 dir_ino, u64 dir_gen)
 {
 	struct rb_node **p = &sctx->orphan_dirs.rb_node;
 	struct rb_node *parent = NULL;
@@ -2868,20 +2869,23 @@ add_orphan_dir_info(struct send_ctx *sctx, u64 dir_ino)
 	while (*p) {
 		parent = *p;
 		entry = rb_entry(parent, struct orphan_dir_info, node);
-		if (dir_ino < entry->ino) {
+		if (dir_ino < entry->ino)
 			p = &(*p)->rb_left;
-		} else if (dir_ino > entry->ino) {
+		else if (dir_ino > entry->ino)
 			p = &(*p)->rb_right;
-		} else {
+		else if (dir_gen < entry->gen)
+			p = &(*p)->rb_left;
+		else if (dir_gen > entry->gen)
+			p = &(*p)->rb_right;
+		else
 			return entry;
-		}
 	}
 
 	odi = kmalloc(sizeof(*odi), GFP_KERNEL);
 	if (!odi)
 		return ERR_PTR(-ENOMEM);
 	odi->ino = dir_ino;
-	odi->gen = 0;
+	odi->gen = dir_gen;
 	odi->last_dir_index_offset = 0;
 
 	rb_link_node(&odi->node, parent, p);
@@ -2889,8 +2893,8 @@ add_orphan_dir_info(struct send_ctx *sctx, u64 dir_ino)
 	return odi;
 }
 
-static struct orphan_dir_info *
-get_orphan_dir_info(struct send_ctx *sctx, u64 dir_ino)
+static struct orphan_dir_info *get_orphan_dir_info(struct send_ctx *sctx,
+						   u64 dir_ino, u64 gen)
 {
 	struct rb_node *n = sctx->orphan_dirs.rb_node;
 	struct orphan_dir_info *entry;
@@ -2901,15 +2905,19 @@ get_orphan_dir_info(struct send_ctx *sctx, u64 dir_ino)
 			n = n->rb_left;
 		else if (dir_ino > entry->ino)
 			n = n->rb_right;
+		else if (gen < entry->gen)
+			n = n->rb_left;
+		else if (gen > entry->gen)
+			n = n->rb_right;
 		else
 			return entry;
 	}
 	return NULL;
 }
 
-static int is_waiting_for_rm(struct send_ctx *sctx, u64 dir_ino)
+static int is_waiting_for_rm(struct send_ctx *sctx, u64 dir_ino, u64 gen)
 {
-	struct orphan_dir_info *odi = get_orphan_dir_info(sctx, dir_ino);
+	struct orphan_dir_info *odi = get_orphan_dir_info(sctx, dir_ino, gen);
 
 	return odi != NULL;
 }
@@ -2954,7 +2962,7 @@ static int can_rmdir(struct send_ctx *sctx, u64 dir, u64 dir_gen,
 	key.type = BTRFS_DIR_INDEX_KEY;
 	key.offset = 0;
 
-	odi = get_orphan_dir_info(sctx, dir);
+	odi = get_orphan_dir_info(sctx, dir, dir_gen);
 	if (odi)
 		key.offset = odi->last_dir_index_offset;
 
@@ -2985,7 +2993,7 @@ static int can_rmdir(struct send_ctx *sctx, u64 dir, u64 dir_gen,
 
 		dm = get_waiting_dir_move(sctx, loc.objectid);
 		if (dm) {
-			odi = add_orphan_dir_info(sctx, dir);
+			odi = add_orphan_dir_info(sctx, dir, dir_gen);
 			if (IS_ERR(odi)) {
 				ret = PTR_ERR(odi);
 				goto out;
@@ -2993,12 +3001,13 @@ static int can_rmdir(struct send_ctx *sctx, u64 dir, u64 dir_gen,
 			odi->gen = dir_gen;
 			odi->last_dir_index_offset = found_key.offset;
 			dm->rmdir_ino = dir;
+			dm->rmdir_gen = dir_gen;
 			ret = 0;
 			goto out;
 		}
 
 		if (loc.objectid > send_progress) {
-			odi = add_orphan_dir_info(sctx, dir);
+			odi = add_orphan_dir_info(sctx, dir, dir_gen);
 			if (IS_ERR(odi)) {
 				ret = PTR_ERR(odi);
 				goto out;
@@ -3038,6 +3047,7 @@ static int add_waiting_dir_move(struct send_ctx *sctx, u64 ino, bool orphanized)
 		return -ENOMEM;
 	dm->ino = ino;
 	dm->rmdir_ino = 0;
+	dm->rmdir_gen = 0;
 	dm->orphanized = orphanized;
 
 	while (*p) {
@@ -3183,7 +3193,7 @@ static int path_loop(struct send_ctx *sctx, struct fs_path *name,
 	while (ino != BTRFS_FIRST_FREE_OBJECTID) {
 		fs_path_reset(name);
 
-		if (is_waiting_for_rm(sctx, ino))
+		if (is_waiting_for_rm(sctx, ino, gen))
 			break;
 		if (is_waiting_for_move(sctx, ino)) {
 			if (*ancestor_ino == 0)
@@ -3223,6 +3233,7 @@ static int apply_dir_move(struct send_ctx *sctx, struct pending_dir_move *pm)
 	u64 parent_ino, parent_gen;
 	struct waiting_dir_move *dm = NULL;
 	u64 rmdir_ino = 0;
+	u64 rmdir_gen;
 	u64 ancestor;
 	bool is_orphan;
 	int ret;
@@ -3237,6 +3248,7 @@ static int apply_dir_move(struct send_ctx *sctx, struct pending_dir_move *pm)
 	dm = get_waiting_dir_move(sctx, pm->ino);
 	ASSERT(dm);
 	rmdir_ino = dm->rmdir_ino;
+	rmdir_gen = dm->rmdir_gen;
 	is_orphan = dm->orphanized;
 	free_waiting_dir_move(sctx, dm);
 
@@ -3273,6 +3285,7 @@ static int apply_dir_move(struct send_ctx *sctx, struct pending_dir_move *pm)
 			dm = get_waiting_dir_move(sctx, pm->ino);
 			ASSERT(dm);
 			dm->rmdir_ino = rmdir_ino;
+			dm->rmdir_gen = rmdir_gen;
 		}
 		goto out;
 	}
@@ -3291,7 +3304,7 @@ static int apply_dir_move(struct send_ctx *sctx, struct pending_dir_move *pm)
 		struct orphan_dir_info *odi;
 		u64 gen;
 
-		odi = get_orphan_dir_info(sctx, rmdir_ino);
+		odi = get_orphan_dir_info(sctx, rmdir_ino, rmdir_gen);
 		if (!odi) {
 			/* already deleted */
 			goto finish;
diff --git a/fs/btrfs/space-info.c b/fs/btrfs/space-info.c
index 64099565ab8f..67e55c5479b8 100644
--- a/fs/btrfs/space-info.c
+++ b/fs/btrfs/space-info.c
@@ -532,7 +532,7 @@ static void shrink_delalloc(struct btrfs_fs_info *fs_info,
 
 	loops = 0;
 	while ((delalloc_bytes || dio_bytes) && loops < 3) {
-		btrfs_start_delalloc_roots(fs_info, items);
+		btrfs_start_delalloc_roots(fs_info, items, true);
 
 		loops++;
 		if (wait_ordered && !trans) {
diff --git a/fs/btrfs/super.c b/fs/btrfs/super.c
index 022f20810089..12d7d3be7cd4 100644
--- a/fs/btrfs/super.c
+++ b/fs/btrfs/super.c
@@ -175,7 +175,7 @@ void __btrfs_handle_fs_error(struct btrfs_fs_info *fs_info, const char *function
 	btrfs_discard_stop(fs_info);
 
 	/* btrfs handle error by forcing the filesystem readonly */
-	sb->s_flags |= SB_RDONLY;
+	btrfs_set_sb_rdonly(sb);
 	btrfs_info(fs_info, "forced readonly");
 	/*
 	 * Note that a running device replace operation is not canceled here
@@ -1953,7 +1953,7 @@ static int btrfs_remount(struct super_block *sb, int *flags, char *data)
 		/* avoid complains from lockdep et al. */
 		up(&fs_info->uuid_tree_rescan_sem);
 
-		sb->s_flags |= SB_RDONLY;
+		btrfs_set_sb_rdonly(sb);
 
 		/*
 		 * Setting SB_RDONLY will put the cleaner thread to
@@ -1964,10 +1964,42 @@ static int btrfs_remount(struct super_block *sb, int *flags, char *data)
 		 */
 		btrfs_delete_unused_bgs(fs_info);
 
+		/*
+		 * The cleaner task could be already running before we set the
+		 * flag BTRFS_FS_STATE_RO (and SB_RDONLY in the superblock).
+		 * We must make sure that after we finish the remount, i.e. after
+		 * we call btrfs_commit_super(), the cleaner can no longer start
+		 * a transaction - either because it was dropping a dead root,
+		 * running delayed iputs or deleting an unused block group (the
+		 * cleaner picked a block group from the list of unused block
+		 * groups before we were able to in the previous call to
+		 * btrfs_delete_unused_bgs()).
+		 */
+		wait_on_bit(&fs_info->flags, BTRFS_FS_CLEANER_RUNNING,
+			    TASK_UNINTERRUPTIBLE);
+
+		/*
+		 * We've set the superblock to RO mode, so we might have made
+		 * the cleaner task sleep without running all pending delayed
+		 * iputs. Go through all the delayed iputs here, so that if an
+		 * unmount happens without remounting RW we don't end up at
+		 * finishing close_ctree() with a non-empty list of delayed
+		 * iputs.
+		 */
+		btrfs_run_delayed_iputs(fs_info);
+
 		btrfs_dev_replace_suspend_for_unmount(fs_info);
 		btrfs_scrub_cancel(fs_info);
 		btrfs_pause_balance(fs_info);
 
+		/*
+		 * Pause the qgroup rescan worker if it is running. We don't want
+		 * it to be still running after we are in RO mode, as after that,
+		 * by the time we unmount, it might have left a transaction open,
+		 * so we would leak the transaction and/or crash.
+		 */
+		btrfs_qgroup_wait_for_completion(fs_info, false);
+
 		ret = btrfs_commit_super(fs_info);
 		if (ret)
 			goto restore;
@@ -2006,7 +2038,7 @@ static int btrfs_remount(struct super_block *sb, int *flags, char *data)
 		if (ret)
 			goto restore;
 
-		sb->s_flags &= ~SB_RDONLY;
+		btrfs_clear_sb_rdonly(sb);
 
 		set_bit(BTRFS_FS_OPEN, &fs_info->flags);
 	}
@@ -2028,6 +2060,8 @@ restore:
 	/* We've hit an error - don't reset SB_RDONLY */
 	if (sb_rdonly(sb))
 		old_flags |= SB_RDONLY;
+	if (!(old_flags & SB_RDONLY))
+		clear_bit(BTRFS_FS_STATE_RO, &fs_info->fs_state);
 	sb->s_flags = old_flags;
 	fs_info->mount_opt = old_opts;
 	fs_info->compress_type = old_compress_type;
diff --git a/fs/btrfs/tests/btrfs-tests.c b/fs/btrfs/tests/btrfs-tests.c
index 8ca334d554af..6bd97bd4cb37 100644
--- a/fs/btrfs/tests/btrfs-tests.c
+++ b/fs/btrfs/tests/btrfs-tests.c
@@ -55,8 +55,14 @@ struct inode *btrfs_new_test_inode(void)
 	struct inode *inode;
 
 	inode = new_inode(test_mnt->mnt_sb);
-	if (inode)
-		inode_init_owner(inode, NULL, S_IFREG);
+	if (!inode)
+		return NULL;
+
+	inode->i_mode = S_IFREG;
+	BTRFS_I(inode)->location.type = BTRFS_INODE_ITEM_KEY;
+	BTRFS_I(inode)->location.objectid = BTRFS_FIRST_FREE_OBJECTID;
+	BTRFS_I(inode)->location.offset = 0;
+	inode_init_owner(inode, NULL, S_IFREG);
 
 	return inode;
 }
diff --git a/fs/btrfs/tests/inode-tests.c b/fs/btrfs/tests/inode-tests.c
index 04022069761d..c9874b12d337 100644
--- a/fs/btrfs/tests/inode-tests.c
+++ b/fs/btrfs/tests/inode-tests.c
@@ -232,11 +232,6 @@ static noinline int test_btrfs_get_extent(u32 sectorsize, u32 nodesize)
 		return ret;
 	}
 
-	inode->i_mode = S_IFREG;
-	BTRFS_I(inode)->location.type = BTRFS_INODE_ITEM_KEY;
-	BTRFS_I(inode)->location.objectid = BTRFS_FIRST_FREE_OBJECTID;
-	BTRFS_I(inode)->location.offset = 0;
-
 	fs_info = btrfs_alloc_dummy_fs_info(nodesize, sectorsize);
 	if (!fs_info) {
 		test_std_err(TEST_ALLOC_FS_INFO);
@@ -835,10 +830,6 @@ static int test_hole_first(u32 sectorsiz