Merge branch 'for-chris-4.11' of git://git.kernel.org/pub/scm/linux/k…

…ernel/git/fdmanana/linux into for-linus-4.11

fs/btrfs/ctree.c

@@ -4159,6 +4159,9 @@ static noinline int push_for_double_split(struct btrfs_trans_handle *trans,
 
 	/* try to push all the items before our slot into the next leaf */
 	slot = path->slots[0];
+	space_needed = data_size;
+	if (slot > 0)
+		space_needed -= btrfs_leaf_free_space(fs_info, path->nodes[0]);
 	ret = push_leaf_left(trans, root, path, 1, space_needed, 0, slot);
 	if (ret < 0)
 		return ret;
@@ -4214,6 +4217,10 @@ static noinline int split_leaf(struct btrfs_trans_handle *trans,
 		if (wret < 0)
 			return wret;
 		if (wret) {
+			space_needed = data_size;
+			if (slot > 0)
+				space_needed -= btrfs_leaf_free_space(fs_info,
+								      l);
 			wret = push_leaf_left(trans, root, path, space_needed,
 					      space_needed, 0, (u32)-1);
 			if (wret < 0)

fs/btrfs/disk-io.c

@@ -2205,11 +2205,9 @@ static void btrfs_stop_all_workers(struct btrfs_fs_info *fs_info)
 	btrfs_destroy_workqueue(fs_info->delalloc_workers);
 	btrfs_destroy_workqueue(fs_info->workers);
 	btrfs_destroy_workqueue(fs_info->endio_workers);
-	btrfs_destroy_workqueue(fs_info->endio_meta_workers);
 	btrfs_destroy_workqueue(fs_info->endio_raid56_workers);
 	btrfs_destroy_workqueue(fs_info->endio_repair_workers);
 	btrfs_destroy_workqueue(fs_info->rmw_workers);
-	btrfs_destroy_workqueue(fs_info->endio_meta_write_workers);
 	btrfs_destroy_workqueue(fs_info->endio_write_workers);
 	btrfs_destroy_workqueue(fs_info->endio_freespace_worker);
 	btrfs_destroy_workqueue(fs_info->submit_workers);
@@ -2219,6 +2217,13 @@ static void btrfs_stop_all_workers(struct btrfs_fs_info *fs_info)
 	btrfs_destroy_workqueue(fs_info->flush_workers);
 	btrfs_destroy_workqueue(fs_info->qgroup_rescan_workers);
 	btrfs_destroy_workqueue(fs_info->extent_workers);
+	/*
+	 * Now that all other work queues are destroyed, we can safely destroy
+	 * the queues used for metadata I/O, since tasks from those other work
+	 * queues can do metadata I/O operations.
+	 */
+	btrfs_destroy_workqueue(fs_info->endio_meta_workers);
+	btrfs_destroy_workqueue(fs_info->endio_meta_write_workers);
 }
 
 static void free_root_extent_buffers(struct btrfs_root *root)
@@ -3261,14 +3266,14 @@ int open_ctree(struct super_block *sb,
 
 fail_block_groups:
 	btrfs_put_block_group_cache(fs_info);
-	btrfs_free_block_groups(fs_info);
 
 fail_tree_roots:
 	free_root_pointers(fs_info, 1);
 	invalidate_inode_pages2(fs_info->btree_inode->i_mapping);
 
 fail_sb_buffer:
 	btrfs_stop_all_workers(fs_info);
+	btrfs_free_block_groups(fs_info);
 fail_alloc:
 fail_iput:
 	btrfs_mapping_tree_free(&fs_info->mapping_tree);
@@ -3977,15 +3982,15 @@ void close_ctree(struct btrfs_fs_info *fs_info)
 
 	btrfs_put_block_group_cache(fs_info);
 
-	btrfs_free_block_groups(fs_info);
-
 	/*
 	 * we must make sure there is not any read request to
 	 * submit after we stopping all workers.
 	 */
 	invalidate_inode_pages2(fs_info->btree_inode->i_mapping);
 	btrfs_stop_all_workers(fs_info);
 
+	btrfs_free_block_groups(fs_info);
+
 	clear_bit(BTRFS_FS_OPEN, &fs_info->flags);
 	free_root_pointers(fs_info, 1);
 

fs/btrfs/extent-tree.c

@@ -9740,6 +9740,11 @@ void btrfs_put_block_group_cache(struct btrfs_fs_info *info)
 	}
 }
 
+/*
+ * Must be called only after stopping all workers, since we could have block
+ * group caching kthreads running, and therefore they could race with us if we
+ * freed the block groups before stopping them.
+ */
 int btrfs_free_block_groups(struct btrfs_fs_info *info)
 {
 	struct btrfs_block_group_cache *block_group;
@@ -9779,9 +9784,6 @@ int btrfs_free_block_groups(struct btrfs_fs_info *info)
 		list_del(&block_group->list);
 		up_write(&block_group->space_info->groups_sem);
 
-		if (block_group->cached == BTRFS_CACHE_STARTED)
-			wait_block_group_cache_done(block_group);
-
 		/*
 		 * We haven't cached this block group, which means we could
 		 * possibly have excluded extents on this block group.
@@ -9791,6 +9793,7 @@ int btrfs_free_block_groups(struct btrfs_fs_info *info)
 			free_excluded_extents(info, block_group);
 
 		btrfs_remove_free_space_cache(block_group);
+		ASSERT(block_group->cached != BTRFS_CACHE_STARTED);
 		ASSERT(list_empty(&block_group->dirty_list));
 		ASSERT(list_empty(&block_group->io_list));
 		ASSERT(list_empty(&block_group->bg_list));

fs/btrfs/file-item.c

@@ -643,7 +643,33 @@ int btrfs_del_csums(struct btrfs_trans_handle *trans,
 
 		/* delete the entire item, it is inside our range */
 		if (key.offset >= bytenr && csum_end <= end_byte) {
-			ret = btrfs_del_item(trans, root, path);
+			int del_nr = 1;
+
+			/*
+			 * Check how many csum items preceding this one in this
+			 * leaf correspond to our range and then delete them all
+			 * at once.
+			 */
+			if (key.offset > bytenr && path->slots[0] > 0) {
+				int slot = path->slots[0] - 1;
+
+				while (slot >= 0) {
+					struct btrfs_key pk;
+
+					btrfs_item_key_to_cpu(leaf, &pk, slot);
+					if (pk.offset < bytenr ||
+					    pk.type != BTRFS_EXTENT_CSUM_KEY ||
+					    pk.objectid !=
+					    BTRFS_EXTENT_CSUM_OBJECTID)
+						break;
+					path->slots[0] = slot;
+					del_nr++;
+					key.offset = pk.offset;
+					slot--;
+				}
+			}
+			ret = btrfs_del_items(trans, root, path,
+					      path->slots[0], del_nr);
 			if (ret)
 				goto out;
 			if (key.offset == bytenr)

fs/btrfs/inode.c

@@ -1331,10 +1331,16 @@ static noinline int run_delalloc_nocow(struct inode *inode,
 			 * either valid or do not exist.
 			 */
 			if (csum_exist_in_range(fs_info, disk_bytenr,
-						num_bytes))
+						num_bytes)) {
+				if (!nolock)
+					btrfs_end_write_no_snapshoting(root);
 				goto out_check;
-			if (!btrfs_inc_nocow_writers(fs_info, disk_bytenr))
+			}
+			if (!btrfs_inc_nocow_writers(fs_info, disk_bytenr)) {
+				if (!nolock)
+					btrfs_end_write_no_snapshoting(root);
 				goto out_check;
+			}
 			nocow = 1;
 		} else if (extent_type == BTRFS_FILE_EXTENT_INLINE) {
 			extent_end = found_key.offset +
@@ -4412,19 +4418,8 @@ int btrfs_truncate_inode_items(struct btrfs_trans_handle *trans,
 		if (found_type > min_type) {
 			del_item = 1;
 		} else {
-			if (item_end < new_size) {
-				/*
-				 * With NO_HOLES mode, for the following mapping
-				 *
-				 * [0-4k][hole][8k-12k]
-				 *
-				 * if truncating isize down to 6k, it ends up
-				 * isize being 8k.
-				 */
-				if (btrfs_fs_incompat(root->fs_info, NO_HOLES))
-					last_size = new_size;
+			if (item_end < new_size)
 				break;
-			}
 			if (found_key.offset >= new_size)
 				del_item = 1;
 			else
@@ -4607,8 +4602,12 @@ int btrfs_truncate_inode_items(struct btrfs_trans_handle *trans,
 			btrfs_abort_transaction(trans, ret);
 	}
 error:
-	if (root->root_key.objectid != BTRFS_TREE_LOG_OBJECTID)
+	if (root->root_key.objectid != BTRFS_TREE_LOG_OBJECTID) {
+		ASSERT(last_size >= new_size);
+		if (!err && last_size > new_size)
+			last_size = new_size;
 		btrfs_ordered_update_i_size(inode, last_size, NULL);
+	}
 
 	btrfs_free_path(path);
 

fs/btrfs/send.c

@@ -1681,6 +1681,9 @@ static int is_inode_existent(struct send_ctx *sctx, u64 ino, u64 gen)
 {
 	int ret;
 
+	if (ino == BTRFS_FIRST_FREE_OBJECTID)
+		return 1;
+
 	ret = get_cur_inode_state(sctx, ino, gen);
 	if (ret < 0)
 		goto out;
@@ -1866,7 +1869,7 @@ static int will_overwrite_ref(struct send_ctx *sctx, u64 dir, u64 dir_gen,
 	 * not deleted and then re-created, if it was then we have no overwrite
 	 * and we can just unlink this entry.
 	 */
-	if (sctx->parent_root) {
+	if (sctx->parent_root && dir != BTRFS_FIRST_FREE_OBJECTID) {
 		ret = get_inode_info(sctx->parent_root, dir, NULL, &gen, NULL,
 				     NULL, NULL, NULL);
 		if (ret < 0 && ret != -ENOENT)
@@ -1934,6 +1937,19 @@ static int did_overwrite_ref(struct send_ctx *sctx,
 	if (ret <= 0)
 		goto out;
 
+	if (dir != BTRFS_FIRST_FREE_OBJECTID) {
+		ret = get_inode_info(sctx->send_root, dir, NULL, &gen, NULL,
+				     NULL, NULL, NULL);
+		if (ret < 0 && ret != -ENOENT)
+			goto out;
+		if (ret) {
+			ret = 0;
+			goto out;
+		}
+		if (gen != dir_gen)
+			goto out;
+	}
+
 	/* check if the ref was overwritten by another ref */
 	ret = lookup_dir_item_inode(sctx->send_root, dir, name, name_len,
 			&ow_inode, &other_type);
@@ -3556,6 +3572,7 @@ static int wait_for_parent_move(struct send_ctx *sctx,
 {
 	int ret = 0;
 	u64 ino = parent_ref->dir;
+	u64 ino_gen = parent_ref->dir_gen;
 	u64 parent_ino_before, parent_ino_after;
 	struct fs_path *path_before = NULL;
 	struct fs_path *path_after = NULL;
@@ -3576,6 +3593,8 @@ static int wait_for_parent_move(struct send_ctx *sctx,
 	 * at get_cur_path()).
 	 */
 	while (ino > BTRFS_FIRST_FREE_OBJECTID) {
+		u64 parent_ino_after_gen;
+
 		if (is_waiting_for_move(sctx, ino)) {
 			/*
 			 * If the current inode is an ancestor of ino in the
@@ -3598,7 +3617,7 @@ static int wait_for_parent_move(struct send_ctx *sctx,
 		fs_path_reset(path_after);
 
 		ret = get_first_ref(sctx->send_root, ino, &parent_ino_after,
-				    NULL, path_after);
+				    &parent_ino_after_gen, path_after);
 		if (ret < 0)
 			goto out;
 		ret = get_first_ref(sctx->parent_root, ino, &parent_ino_before,
@@ -3615,10 +3634,20 @@ static int wait_for_parent_move(struct send_ctx *sctx,
 		if (ino > sctx->cur_ino &&
 		    (parent_ino_before != parent_ino_after || len1 != len2 ||
 		     memcmp(path_before->start, path_after->start, len1))) {
-			ret = 1;
-			break;
+			u64 parent_ino_gen;
+
+			ret = get_inode_info(sctx->parent_root, ino, NULL,
+					     &parent_ino_gen, NULL, NULL, NULL,
+					     NULL);
+			if (ret < 0)
+				goto out;
+			if (ino_gen == parent_ino_gen) {
+				ret = 1;
+				break;
+			}
 		}
 		ino = parent_ino_after;
+		ino_gen = parent_ino_after_gen;
 	}
 
 out:
@@ -5277,6 +5306,81 @@ static int get_last_extent(struct send_ctx *sctx, u64 offset)
 	return ret;
 }
 
+static int range_is_hole_in_parent(struct send_ctx *sctx,
+				   const u64 start,
+				   const u64 end)
+{
+	struct btrfs_path *path;
+	struct btrfs_key key;
+	struct btrfs_root *root = sctx->parent_root;
+	u64 search_start = start;
+	int ret;
+
+	path = alloc_path_for_send();
+	if (!path)
+		return -ENOMEM;
+
+	key.objectid = sctx->cur_ino;
+	key.type = BTRFS_EXTENT_DATA_KEY;
+	key.offset = search_start;
+	ret = btrfs_search_slot(NULL, root, &key, path, 0, 0);
+	if (ret < 0)
+		goto out;
+	if (ret > 0 && path->slots[0] > 0)
+		path->slots[0]--;
+
+	while (search_start < end) {
+		struct extent_buffer *leaf = path->nodes[0];
+		int slot = path->slots[0];
+		struct btrfs_file_extent_item *fi;
+		u64 extent_end;
+
+		if (slot >= btrfs_header_nritems(leaf)) {
+			ret = btrfs_next_leaf(root, path);
+			if (ret < 0)
+				goto out;
+			else if (ret > 0)
+				break;
+			continue;
+		}
+
+		btrfs_item_key_to_cpu(leaf, &key, slot);
+		if (key.objectid < sctx->cur_ino ||
+		    key.type < BTRFS_EXTENT_DATA_KEY)
+			goto next;
+		if (key.objectid > sctx->cur_ino ||
+		    key.type > BTRFS_EXTENT_DATA_KEY ||
+		    key.offset >= end)
+			break;
+
+		fi = btrfs_item_ptr(leaf, slot, struct btrfs_file_extent_item);
+		if (btrfs_file_extent_type(leaf, fi) ==
+		    BTRFS_FILE_EXTENT_INLINE) {
+			u64 size = btrfs_file_extent_inline_len(leaf, slot, fi);
+
+			extent_end = ALIGN(key.offset + size,
+					   root->fs_info->sectorsize);
+		} else {
+			extent_end = key.offset +
+				btrfs_file_extent_num_bytes(leaf, fi);
+		}
+		if (extent_end <= start)
+			goto next;
+		if (btrfs_file_extent_disk_bytenr(leaf, fi) == 0) {
+			search_start = extent_end;
+			goto next;
+		}
+		ret = 0;
+		goto out;
+next:
+		path->slots[0]++;
+	}
+	ret = 1;
+out:
+	btrfs_free_path(path);
+	return ret;
+}
+
 static int maybe_send_hole(struct send_ctx *sctx, struct btrfs_path *path,
 			   struct btrfs_key *key)
 {
@@ -5321,8 +5425,17 @@ static int maybe_send_hole(struct send_ctx *sctx, struct btrfs_path *path,
 			return ret;
 	}
 
-	if (sctx->cur_inode_last_extent < key->offset)
-		ret = send_hole(sctx, key->offset);
+	if (sctx->cur_inode_last_extent < key->offset) {
+		ret = range_is_hole_in_parent(sctx,
+					      sctx->cur_inode_last_extent,
+					      key->offset);
+		if (ret < 0)
+			return ret;
+		else if (ret == 0)
+			ret = send_hole(sctx, key->offset);
+		else
+			ret = 0;
+	}
 	sctx->cur_inode_last_extent = extent_end;
 	return ret;
 }