From: Andreas Dilger Subject: Re: ext2fs_block_iterate() on fast symlink Date: Thu, 21 Jun 2007 03:33:43 -0600 Message-ID: <20070621093343.GO5181@schatzie.adilger.int> References: <20070620125653.GG27218@duck.suse.cz> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Cc: linux-ext4@vger.kernel.org, tytso@mit.edu To: Jan Kara Return-path: Received: from mail.clusterfs.com ([206.168.112.78]:50256 "EHLO mail.clusterfs.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1751218AbXFUJdq (ORCPT ); Thu, 21 Jun 2007 05:33:46 -0400 Content-Disposition: inline In-Reply-To: <20070620125653.GG27218@duck.suse.cz> Sender: linux-ext4-owner@vger.kernel.org List-Id: linux-ext4.vger.kernel.org On Jun 20, 2007 14:56 +0200, Jan Kara wrote: > when ext2fs_block_iterate() is called on a fast symlink (and I assume > device inodes would be no different), then random things happen - the > problem is ext2fs_block_iterate() just blindly takes portions of the inode > and treats them as block numbers. Now I agree that garbage went in (it > makes no sence to call this function on such inode) so garbage results but > maybe it would be nicer to handle it more gracefully. Attached patch should > do it. > --- a/lib/ext2fs/inode.c 2007-06-20 13:55:52.000000000 +0200 > +++ b/lib/ext2fs/inode.c 2007-06-20 14:11:15.000000000 +0200 > @@ -771,6 +771,10 @@ errcode_t ext2fs_get_blocks(ext2_filsys > retval = ext2fs_read_inode(fs, ino, &inode); > if (retval) > return retval; > + if (LINUX_S_ISCHR(inode.i_mode) || LINUX_S_ISBLK(inode.i_mode) || > + (LINUX_S_ISLNK(inode.i_mode) && > + ext2fs_inode_data_blocks(fs, &inode) == 0)) > + return EXT2_ET_INVAL_INODE_TYPE; I would prefer that we NOT continue to make fast symlinks conditional upon the i_blocks count. That causes problems if e.g. an EA block is present (that would cause this blocks == 0 test to incorrectly fail), and may making the check (blocks - !!i_file_acl) can still fail for other reasons where a block is added to an inode (e.g. if we have larger EAs, etc). I'd prefer to make this check "i_size < sizeof(i_block)" or similar, which has always been true for fast symlinks, for every kernel that I have ever seen. Cheers, Andreas -- Andreas Dilger Principal Software Engineer Cluster File Systems, Inc.