1 files changed, 270 insertions, 78 deletions

diff --git a/fs/btrfs/tree-log.c b/fs/btrfs/tree-log.c
index 0dba09334a16..f06454a55e00 100644
--- a/fs/btrfs/tree-log.c
+++ b/fs/btrfs/tree-log.c
@@ -27,6 +27,7 @@
 #define LOG_INODE_ALL 0
 #define LOG_INODE_EXISTS 1
 #define LOG_OTHER_INODE 2
+#define LOG_OTHER_INODE_ALL 3
 
 /*
  * directory trouble cases
@@ -1144,7 +1145,7 @@ next:
 	}
 	btrfs_release_path(path);
 
-	/* look for a conflicing name */
+	/* look for a conflicting name */
 	di = btrfs_lookup_dir_item(trans, root, path, btrfs_ino(dir),
 				   name, namelen, 0);
 	if (di && !IS_ERR(di)) {
@@ -1330,6 +1331,67 @@ out:
 	return ret;
 }
 
+static int add_link(struct btrfs_trans_handle *trans, struct btrfs_root *root,
+		    struct inode *dir, struct inode *inode, const char *name,
+		    int namelen, u64 ref_index)
+{
+	struct btrfs_dir_item *dir_item;
+	struct btrfs_key key;
+	struct btrfs_path *path;
+	struct inode *other_inode = NULL;
+	int ret;
+
+	path = btrfs_alloc_path();
+	if (!path)
+		return -ENOMEM;
+
+	dir_item = btrfs_lookup_dir_item(NULL, root, path,
+					 btrfs_ino(BTRFS_I(dir)),
+					 name, namelen, 0);
+	if (!dir_item) {
+		btrfs_release_path(path);
+		goto add_link;
+	} else if (IS_ERR(dir_item)) {
+		ret = PTR_ERR(dir_item);
+		goto out;
+	}
+
+	/*
+	 * Our inode's dentry collides with the dentry of another inode which is
+	 * in the log but not yet processed since it has a higher inode number.
+	 * So delete that other dentry.
+	 */
+	btrfs_dir_item_key_to_cpu(path->nodes[0], dir_item, &key);
+	btrfs_release_path(path);
+	other_inode = read_one_inode(root, key.objectid);
+	if (!other_inode) {
+		ret = -ENOENT;
+		goto out;
+	}
+	ret = btrfs_unlink_inode(trans, root, BTRFS_I(dir), BTRFS_I(other_inode),
+				 name, namelen);
+	if (ret)
+		goto out;
+	/*
+	 * If we dropped the link count to 0, bump it so that later the iput()
+	 * on the inode will not free it. We will fixup the link count later.
+	 */
+	if (other_inode->i_nlink == 0)
+		inc_nlink(other_inode);
+
+	ret = btrfs_run_delayed_items(trans);
+	if (ret)
+		goto out;
+add_link:
+	ret = btrfs_add_link(trans, BTRFS_I(dir), BTRFS_I(inode),
+			     name, namelen, 0, ref_index);
+out:
+	iput(other_inode);
+	btrfs_free_path(path);
+
+	return ret;
+}
+
 /*
  * replay one inode back reference item found in the log tree.
  * eb, slot and key refer to the buffer and key found in the log tree.
@@ -1466,9 +1528,8 @@ static noinline int add_inode_ref(struct btrfs_trans_handle *trans,
 				goto out;
 
 			/* insert our name */
-			ret = btrfs_add_link(trans, BTRFS_I(dir),
-					BTRFS_I(inode),
-					name, namelen, 0, ref_index);
+			ret = add_link(trans, root, dir, inode, name, namelen,
+				       ref_index);
 			if (ret)
 				goto out;
 
@@ -2663,7 +2724,7 @@ static noinline int walk_down_log_tree(struct btrfs_trans_handle *trans,
 
 				if (trans) {
 					btrfs_tree_lock(next);
-					btrfs_set_lock_blocking(next);
+					btrfs_set_lock_blocking_write(next);
 					clean_tree_block(fs_info, next);
 					btrfs_wait_tree_block_writeback(next);
 					btrfs_tree_unlock(next);
@@ -2747,7 +2808,7 @@ static noinline int walk_up_log_tree(struct btrfs_trans_handle *trans,
 
 				if (trans) {
 					btrfs_tree_lock(next);
-					btrfs_set_lock_blocking(next);
+					btrfs_set_lock_blocking_write(next);
 					clean_tree_block(fs_info, next);
 					btrfs_wait_tree_block_writeback(next);
 					btrfs_tree_unlock(next);
@@ -2829,7 +2890,7 @@ static int walk_log_tree(struct btrfs_trans_handle *trans,
 
 			if (trans) {
 				btrfs_tree_lock(next);
-				btrfs_set_lock_blocking(next);
+				btrfs_set_lock_blocking_write(next);
 				clean_tree_block(fs_info, next);
 				btrfs_wait_tree_block_writeback(next);
 				btrfs_tree_unlock(next);
@@ -3149,7 +3210,7 @@ int btrfs_sync_log(struct btrfs_trans_handle *trans,
 	mutex_unlock(&log_root_tree->log_mutex);
 
 	/*
-	 * nobody else is going to jump in and write the the ctree
+	 * Nobody else is going to jump in and write the ctree
 	 * super here because the log_commit atomic below is protecting
 	 * us.  We must be called with a transaction handle pinning
 	 * the running transaction open, so a full commit can't hop
@@ -3201,8 +3262,6 @@ static void free_log_tree(struct btrfs_trans_handle *trans,
 			  struct btrfs_root *log)
 {
 	int ret;
-	u64 start;
-	u64 end;
 	struct walk_control wc = {
 		.free = 1,
 		.process_func = process_one_buffer
@@ -3216,18 +3275,8 @@ static void free_log_tree(struct btrfs_trans_handle *trans,
 			btrfs_handle_fs_error(log->fs_info, ret, NULL);
 	}
 
-	while (1) {
-		ret = find_first_extent_bit(&log->dirty_log_pages,
-				0, &start, &end,
-				EXTENT_DIRTY | EXTENT_NEW | EXTENT_NEED_WAIT,
-				NULL);
-		if (ret)
-			break;
-
-		clear_extent_bits(&log->dirty_log_pages, start, end,
-				  EXTENT_DIRTY | EXTENT_NEW | EXTENT_NEED_WAIT);
-	}
-
+	clear_extent_bits(&log->dirty_log_pages, 0, (u64)-1,
+			  EXTENT_DIRTY | EXTENT_NEW | EXTENT_NEED_WAIT);
 	free_extent_buffer(log->node);
 	kfree(log);
 }
@@ -3718,6 +3767,8 @@ static int drop_objectid_items(struct btrfs_trans_handle *trans,
 		found_key.type = 0;
 		ret = btrfs_bin_search(path->nodes[0], &found_key, 0,
 				       &start_slot);
+		if (ret < 0)
+			break;
 
 		ret = btrfs_del_items(trans, log, path, start_slot,
 				      path->slots[0] - start_slot + 1);
@@ -4383,20 +4434,33 @@ static int btrfs_log_changed_extents(struct btrfs_trans_handle *trans,
 	struct extent_map *em, *n;
 	struct list_head extents;
 	struct extent_map_tree *tree = &inode->extent_tree;
-	u64 logged_start, logged_end;
 	u64 test_gen;
 	int ret = 0;
 	int num = 0;
 
 	INIT_LIST_HEAD(&extents);
 
-	down_write(&inode->dio_sem);
 	write_lock(&tree->lock);
 	test_gen = root->fs_info->last_trans_committed;
-	logged_start = start;
-	logged_end = end;
 
 	list_for_each_entry_safe(em, n, &tree->modified_extents, list) {
+		/*
+		 * Skip extents outside our logging range. It's important to do
+		 * it for correctness because if we don't ignore them, we may
+		 * log them before their ordered extent completes, and therefore
+		 * we could log them without logging their respective checksums
+		 * (the checksum items are added to the csum tree at the very
+		 * end of btrfs_finish_ordered_io()). Also leave such extents
+		 * outside of our range in the list, since we may have another
+		 * ranged fsync in the near future that needs them. If an extent
+		 * outside our range corresponds to a hole, log it to avoid
+		 * leaving gaps between extents (fsck will complain when we are
+		 * not using the NO_HOLES feature).
+		 */
+		if ((em->start > end || em->start + em->len <= start) &&
+		    em->block_start != EXTENT_MAP_HOLE)
+			continue;
+
 		list_del_init(&em->list);
 		/*
 		 * Just an arbitrary number, this can be really CPU intensive
@@ -4418,11 +4482,6 @@ static int btrfs_log_changed_extents(struct btrfs_trans_handle *trans,
 		    em->start >= i_size_read(&inode->vfs_inode))
 			continue;
 
-		if (em->start < logged_start)
-			logged_start = em->start;
-		if ((em->start + em->len - 1) > logged_end)
-			logged_end = em->start + em->len - 1;
-
 		/* Need a ref to keep it from getting evicted from cache */
 		refcount_inc(&em->refs);
 		set_bit(EXTENT_FLAG_LOGGING, &em->flags);
@@ -4456,7 +4515,6 @@ process:
 	}
 	WARN_ON(!list_empty(&extents));
 	write_unlock(&tree->lock);
-	up_write(&inode->dio_sem);
 
 	btrfs_release_path(path);
 	if (!ret)
@@ -4652,7 +4710,8 @@ static int btrfs_log_trailing_hole(struct btrfs_trans_handle *trans,
 			ASSERT(len == i_size ||
 			       (len == fs_info->sectorsize &&
 				btrfs_file_extent_compression(leaf, extent) !=
-				BTRFS_COMPRESS_NONE));
+				BTRFS_COMPRESS_NONE) ||
+			       (len < i_size && i_size < fs_info->sectorsize));
 			return 0;
 		}
 
@@ -4721,7 +4780,7 @@ static int btrfs_check_ref_name_override(struct extent_buffer *eb,
 					 const int slot,
 					 const struct btrfs_key *key,
 					 struct btrfs_inode *inode,
-					 u64 *other_ino)
+					 u64 *other_ino, u64 *other_parent)
 {
 	int ret;
 	struct btrfs_path *search_path;
@@ -4784,8 +4843,13 @@ static int btrfs_check_ref_name_override(struct extent_buffer *eb,
 			btrfs_dir_item_key_to_cpu(search_path->nodes[0],
 						  di, &di_key);
 			if (di_key.type == BTRFS_INODE_ITEM_KEY) {
-				ret = 1;
-				*other_ino = di_key.objectid;
+				if (di_key.objectid != key->objectid) {
+					ret = 1;
+					*other_ino = di_key.objectid;
+					*other_parent = parent;
+				} else {
+					ret = 0;
+				}
 			} else {
 				ret = -EAGAIN;
 			}
@@ -4805,6 +4869,144 @@ out:
 	return ret;
 }
 
+struct btrfs_ino_list {
+	u64 ino;
+	u64 parent;
+	struct list_head list;
+};
+
+static int log_conflicting_inodes(struct btrfs_trans_handle *trans,
+				  struct btrfs_root *root,
+				  struct btrfs_path *path,
+				  struct btrfs_log_ctx *ctx,
+				  u64 ino, u64 parent)
+{
+	struct btrfs_ino_list *ino_elem;
+	LIST_HEAD(inode_list);
+	int ret = 0;
+
+	ino_elem = kmalloc(sizeof(*ino_elem), GFP_NOFS);
+	if (!ino_elem)
+		return -ENOMEM;
+	ino_elem->ino = ino;
+	ino_elem->parent = parent;
+	list_add_tail(&ino_elem->list, &inode_list);
+
+	while (!list_empty(&inode_list)) {
+		struct btrfs_fs_info *fs_info = root->fs_info;
+		struct btrfs_key key;
+		struct inode *inode;
+
+		ino_elem = list_first_entry(&inode_list, struct btrfs_ino_list,
+					    list);
+		ino = ino_elem->ino;
+		parent = ino_elem->parent;
+		list_del(&ino_elem->list);
+		kfree(ino_elem);
+		if (ret)
+			continue;
+
+		btrfs_release_path(path);
+
+		key.objectid = ino;
+		key.type = BTRFS_INODE_ITEM_KEY;
+		key.offset = 0;
+		inode = btrfs_iget(fs_info->sb, &key, root, NULL);
+		/*
+		 * If the other inode that had a conflicting dir entry was
+		 * deleted in the current transaction, we need to log its parent
+		 * directory.
+		 */
+		if (IS_ERR(inode)) {
+			ret = PTR_ERR(inode);
+			if (ret == -ENOENT) {
+				key.objectid = parent;
+				inode = btrfs_iget(fs_info->sb, &key, root,
+						   NULL);
+				if (IS_ERR(inode)) {
+					ret = PTR_ERR(inode);
+				} else {
+					ret = btrfs_log_inode(trans, root,
+						      BTRFS_I(inode),
+						      LOG_OTHER_INODE_ALL,
+						      0, LLONG_MAX, ctx);
+					iput(inode);
+				}
+			}
+			continue;
+		}
+		/*
+		 * We are safe logging the other inode without acquiring its
+		 * lock as long as we log with the LOG_INODE_EXISTS mode. We
+		 * are safe against concurrent renames of the other inode as
+		 * well because during a rename we pin the log and update the
+		 * log with the new name before we unpin it.
+		 */
+		ret = btrfs_log_inode(trans, root, BTRFS_I(inode),
+				      LOG_OTHER_INODE, 0, LLONG_MAX, ctx);
+		if (ret) {
+			iput(inode);
+			continue;
+		}
+
+		key.objectid = ino;
+		key.type = BTRFS_INODE_REF_KEY;
+		key.offset = 0;
+		ret = btrfs_search_slot(NULL, root, &key, path, 0, 0);
+		if (ret < 0) {
+			iput(inode);
+			continue;
+		}
+
+		while (true) {
+			struct extent_buffer *leaf = path->nodes[0];
+			int slot = path->slots[0];
+			u64 other_ino = 0;
+			u64 other_parent = 0;
+
+			if (slot >= btrfs_header_nritems(leaf)) {
+				ret = btrfs_next_leaf(root, path);
+				if (ret < 0) {
+					break;
+				} else if (ret > 0) {
+					ret = 0;
+					break;
+				}
+				continue;
+			}
+
+			btrfs_item_key_to_cpu(leaf, &key, slot);
+			if (key.objectid != ino ||
+			    (key.type != BTRFS_INODE_REF_KEY &&
+			     key.type != BTRFS_INODE_EXTREF_KEY)) {
+				ret = 0;
+				break;
+			}
+
+			ret = btrfs_check_ref_name_override(leaf, slot, &key,
+					BTRFS_I(inode), &other_ino,
+					&other_parent);
+			if (ret < 0)
+				break;
+			if (ret > 0) {
+				ino_elem = kmalloc(sizeof(*ino_elem), GFP_NOFS);
+				if (!ino_elem) {
+					ret = -ENOMEM;
+					break;
+				}
+				ino_elem->ino = other_ino;
+				ino_elem->parent = other_parent;
+				list_add_tail(&ino_elem->list, &inode_list);
+				ret = 0;
+			}
+			path->slots[0]++;
+		}
+		iput(inode);
+	}
+
+	return ret;
+}
+
 /* log a single inode in the tree log.
  * At least one parent directory for this inode must exist in the tree
  * or be logged already.
@@ -4844,6 +5046,7 @@ static int btrfs_log_inode(struct btrfs_trans_handle *trans,
 	u64 logged_isize = 0;
 	bool need_log_inode_item = true;
 	bool xattrs_logged = false;
+	bool recursive_logging = false;
 
 	path = btrfs_alloc_path();
 	if (!path)
@@ -4889,8 +5092,12 @@ static int btrfs_log_inode(struct btrfs_trans_handle *trans,
 		return ret;
 	}
 
-	if (inode_only == LOG_OTHER_INODE) {
-		inode_only = LOG_INODE_EXISTS;
+	if (inode_only == LOG_OTHER_INODE || inode_only == LOG_OTHER_INODE_ALL) {
+		recursive_logging = true;
+		if (inode_only == LOG_OTHER_INODE)
+			inode_only = LOG_INODE_EXISTS;
+		else
+			inode_only = LOG_INODE_ALL;
 		mutex_lock_nested(&inode->log_mutex, SINGLE_DEPTH_NESTING);
 	} else {
 		mutex_lock(&inode->log_mutex);
@@ -4985,20 +5192,19 @@ again:
 
 		if ((min_key.type == BTRFS_INODE_REF_KEY ||
 		     min_key.type == BTRFS_INODE_EXTREF_KEY) &&
-		    inode->generation == trans->transid) {
+		    inode->generation == trans->transid &&
+		    !recursive_logging) {
 			u64 other_ino = 0;
+			u64 other_parent = 0;
 
 			ret = btrfs_check_ref_name_override(path->nodes[0],
 					path->slots[0], &min_key, inode,
-					&other_ino);
+					&other_ino, &other_parent);
 			if (ret < 0) {
 				err = ret;
 				goto out_unlock;
 			} else if (ret > 0 && ctx &&
 				   other_ino != btrfs_ino(BTRFS_I(ctx->inode))) {
-				struct btrfs_key inode_key;
-				struct inode *other_inode;
-
 				if (ins_nr > 0) {
 					ins_nr++;
 				} else {
@@ -5014,43 +5220,13 @@ again:
 					goto out_unlock;
 				}
 				ins_nr = 0;
-				btrfs_release_path(path);
-				inode_key.objectid = other_ino;
-				inode_key.type = BTRFS_INODE_ITEM_KEY;
-				inode_key.offset = 0;
-				other_inode = btrfs_iget(fs_info->sb,
-							 &inode_key, root,
-							 NULL);
-				/*
-				 * If the other inode that had a conflicting dir
-				 * entry was deleted in the current transaction,
-				 * we don't need to do more work nor fallback to
-				 * a transaction commit.
-				 */
-				if (other_inode == ERR_PTR(-ENOENT)) {
-					goto next_key;
-				} else if (IS_ERR(other_inode)) {
-					err = PTR_ERR(other_inode);
-					goto out_unlock;
-				}
-				/*
-				 * We are safe logging the other inode without
-				 * acquiring its i_mutex as long as we log with
-				 * the LOG_INODE_EXISTS mode. We're safe against
-				 * concurrent renames of the other inode as well
-				 * because during a rename we pin the log and
-				 * update the log with the new name before we
-				 * unpin it.
-				 */
-				err = btrfs_log_inode(trans, root,
-						BTRFS_I(other_inode),
-						LOG_OTHER_INODE, 0, LLONG_MAX,
-						ctx);
-				iput(other_inode);
+
+				err = log_conflicting_inodes(trans, root, path,
+						ctx, other_ino, other_parent);
 				if (err)
 					goto out_unlock;
-				else
-					goto next_key;
+				btrfs_release_path(path);
+				goto next_key;
 			}
 		}
 
@@ -5762,6 +5938,22 @@ static int btrfs_log_inode_parent(struct btrfs_trans_handle *trans,
 			goto end_trans;
 	}
 
+	/*
+	 * If a new hard link was added to the inode in the current transaction
+	 * and its link count is now greater than 1, we need to fallback to a
+	 * transaction commit, otherwise we can end up not logging all its new
+	 * parents for all the hard links. Here just from the dentry used to
+	 * fsync, we can not visit the ancestor inodes for all the other hard
+	 * links to figure out if any is new, so we fallback to a transaction
+	 * commit (instead of adding a lot of complexity of scanning a btree,
+	 * since this scenario is not a common use case).
+	 */
+	if (inode->vfs_inode.i_nlink > 1 &&
+	    inode->last_link_trans > last_committed) {
+		ret = -EMLINK;
+		goto end_trans;
+	}
+
 	while (1) {
 		if (!parent || d_really_is_negative(parent) || sb != parent->d_sb)
 			break;