From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from cn.fujitsu.com ([222.73.24.84]:33431 "EHLO song.cn.fujitsu.com" rhost-flags-OK-FAIL-OK-OK) by vger.kernel.org with ESMTP id S1755776AbcLOJFt (ORCPT ); Thu, 15 Dec 2016 04:05:49 -0500 From: Qu Wenruo To: linux-btrfs@vger.kernel.org, dsterba@suse.cz Cc: t-itoh@jp.fujitsu.com Subject: [PATCH v2 1/3] btrfs-progs: file-item: Fix wrong file extents inserted Date: Thu, 15 Dec 2016 17:03:29 +0800 Message-Id: <20161215090331.27486-2-quwenruo@cn.fujitsu.com> In-Reply-To: <20161215090331.27486-1-quwenruo@cn.fujitsu.com> References: <20161215090331.27486-1-quwenruo@cn.fujitsu.com> MIME-Version: 1.0 Sender: linux-btrfs-owner@vger.kernel.org List-ID: If we specify NO_HOLES incompat feature when converting, the result image still uses hole file extents. And further more, the hole is incorrect as its disk_num_bytes is not zero. The problem is at btrfs_insert_file_extent() which doesn't check if we are going to insert hole file extent. Modify it to skip hole file extents to allow it follow restrict NO_HOLES flag. And since no_holes flag can be triggered on half-way, so current fsck won't report such error, as it consider it as old file holes. Signed-off-by: Qu Wenruo --- convert/main.c | 2 +- file-item.c | 11 +++++++++++ 2 files changed, 12 insertions(+), 1 deletion(-) diff --git a/convert/main.c b/convert/main.c index fd6f77b..15f14af 100644 --- a/convert/main.c +++ b/convert/main.c @@ -336,7 +336,7 @@ static int record_file_blocks(struct blk_iterate_data *data, key.offset > cur_off); fi = btrfs_item_ptr(node, slot, struct btrfs_file_extent_item); extent_disk_bytenr = btrfs_file_extent_disk_bytenr(node, fi); - extent_num_bytes = btrfs_file_extent_disk_num_bytes(node, fi); + extent_num_bytes = btrfs_file_extent_num_bytes(node, fi); BUG_ON(cur_off - key.offset >= extent_num_bytes); btrfs_release_path(&path); diff --git a/file-item.c b/file-item.c index 67c0b4f..e462b4b 100644 --- a/file-item.c +++ b/file-item.c @@ -36,11 +36,22 @@ int btrfs_insert_file_extent(struct btrfs_trans_handle *trans, u64 disk_num_bytes, u64 num_bytes) { int ret = 0; + int is_hole = 0; struct btrfs_file_extent_item *item; struct btrfs_key file_key; struct btrfs_path *path; struct extent_buffer *leaf; + if (offset == 0) + is_hole = 1; + /* For NO_HOLES, we don't insert hole file extent */ + if (btrfs_fs_incompat(root->fs_info, NO_HOLES) && is_hole) + return 0; + + /* For hole, its disk_bytenr and disk_num_bytes must be 0 */ + if (is_hole) + disk_num_bytes = 0; + path = btrfs_alloc_path(); if (!path) return -ENOMEM; -- 2.10.2