From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1757201Ab2ILXuP (ORCPT ); Wed, 12 Sep 2012 19:50:15 -0400 Received: from mail-oa0-f46.google.com ([209.85.219.46]:58620 "EHLO mail-oa0-f46.google.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1757007Ab2ILXof (ORCPT ); Wed, 12 Sep 2012 19:44:35 -0400 From: Greg Kroah-Hartman To: linux-kernel@vger.kernel.org, stable@vger.kernel.org Cc: Greg KH , Kristian Nielsen , Jan Kara Subject: [ 40/46] ext3: Fix fdatasync() for files with only i_size changes Date: Wed, 12 Sep 2012 16:39:30 -0700 Message-Id: <20120912233821.857578372@linuxfoundation.org> X-Mailer: git-send-email 1.7.10.1.362.g242cab3 In-Reply-To: <20120912233817.662663809@linuxfoundation.org> References: <20120912233817.662663809@linuxfoundation.org> User-Agent: quilt/0.60-2.1.2 Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org From: Greg KH 3.0-stable review patch. If anyone has any objections, please let me know. ------------------ From: Jan Kara commit 156bddd8e505b295540f3ca0e27dda68cb0d49aa upstream. Code tracking when transaction needs to be committed on fdatasync(2) forgets to handle a situation when only inode's i_size is changed. Thus in such situations fdatasync(2) doesn't force transaction with new i_size to disk and that can result in wrong i_size after a crash. Fix the issue by updating inode's i_datasync_tid whenever its size is updated. Reported-by: Kristian Nielsen Signed-off-by: Jan Kara Signed-off-by: Greg Kroah-Hartman --- fs/ext3/inode.c | 17 ++++++++++++++--- 1 file changed, 14 insertions(+), 3 deletions(-) --- a/fs/ext3/inode.c +++ b/fs/ext3/inode.c @@ -3013,6 +3013,8 @@ static int ext3_do_update_inode(handle_t struct ext3_inode_info *ei = EXT3_I(inode); struct buffer_head *bh = iloc->bh; int err = 0, rc, block; + int need_datasync = 0; + __le32 disksize; again: /* we can't allow multiple procs in here at once, its a bit racey */ @@ -3050,7 +3052,11 @@ again: raw_inode->i_gid_high = 0; } raw_inode->i_links_count = cpu_to_le16(inode->i_nlink); - raw_inode->i_size = cpu_to_le32(ei->i_disksize); + disksize = cpu_to_le32(ei->i_disksize); + if (disksize != raw_inode->i_size) { + need_datasync = 1; + raw_inode->i_size = disksize; + } raw_inode->i_atime = cpu_to_le32(inode->i_atime.tv_sec); raw_inode->i_ctime = cpu_to_le32(inode->i_ctime.tv_sec); raw_inode->i_mtime = cpu_to_le32(inode->i_mtime.tv_sec); @@ -3066,8 +3072,11 @@ again: if (!S_ISREG(inode->i_mode)) { raw_inode->i_dir_acl = cpu_to_le32(ei->i_dir_acl); } else { - raw_inode->i_size_high = - cpu_to_le32(ei->i_disksize >> 32); + disksize = cpu_to_le32(ei->i_disksize >> 32); + if (disksize != raw_inode->i_size_high) { + raw_inode->i_size_high = disksize; + need_datasync = 1; + } if (ei->i_disksize > 0x7fffffffULL) { struct super_block *sb = inode->i_sb; if (!EXT3_HAS_RO_COMPAT_FEATURE(sb, @@ -3120,6 +3129,8 @@ again: ext3_clear_inode_state(inode, EXT3_STATE_NEW); atomic_set(&ei->i_sync_tid, handle->h_transaction->t_tid); + if (need_datasync) + atomic_set(&ei->i_datasync_tid, handle->h_transaction->t_tid); out_brelse: brelse (bh); ext3_std_error(inode->i_sb, err);