From patchwork Wed Nov 22 19:36:44 2023 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: Al Viro X-Patchwork-Id: 13465415 Authentication-Results: smtp.subspace.kernel.org; dkim=pass (2048-bit key) header.d=linux.org.uk header.i=@linux.org.uk header.b="K/Tc2GyY" Received: from zeniv.linux.org.uk (zeniv.linux.org.uk [IPv6:2a03:a000:7:0:5054:ff:fe1c:15ff]) by lindbergh.monkeyblade.net (Postfix) with ESMTPS id 989421A4; Wed, 22 Nov 2023 11:36:54 -0800 (PST) DKIM-Signature: v=1; a=rsa-sha256; q=dns/txt; c=relaxed/relaxed; d=linux.org.uk; s=zeniv-20220401; h=Sender:Content-Transfer-Encoding: MIME-Version:References:In-Reply-To:Message-Id:Date:Subject:Cc:To:From: Reply-To:Content-Type:Content-ID:Content-Description; bh=BlfFBoYl8WljiWMPq8jOBx3qNB8G8Xt5J/at8Fwypa8=; b=K/Tc2GyYjNHRjlY66jCcTKv40h uI2eHqgIdzYpfZ8iPDPnzytxzUnw8C0mRdemI1USPVe0uuygI28Nzjm0arhCOluRn6lc6XrntEs5+ rUK7HHMr9xc2gQKMIhJ+WIuq8K8TBbFRtgMI45u6VyMP5TIWWxzdjA+r1X1NRGOpr5JpWWBdMa4ZW dJjn6DdaHNIs/wqMAfvUz9aa5VbLxEaeR18lZi+Q7x6NHo/YNOu9ArS3vGuSQTQXgsb7izY0ElA7v jLnL47vLk9zswGvFOm99fDOeT4GbsXJikaM48F7HxYvVntvW9MxwZJVQL6HePppULdL35ahLasKKf lFDbKPhw==; Received: from viro by zeniv.linux.org.uk with local (Exim 4.96 #2 (Red Hat Linux)) id 1r5t1o-001l1e-1Y; Wed, 22 Nov 2023 19:36:52 +0000 From: Al Viro To: linux-fsdevel@vger.kernel.org Cc: Linus Torvalds , Mo Zou , Jan Kara , linux-kernel@vger.kernel.org Subject: [PATCH 1/9] reiserfs: Avoid touching renamed directory if parent does not change Date: Wed, 22 Nov 2023 19:36:44 +0000 Message-Id: <20231122193652.419091-1-viro@zeniv.linux.org.uk> X-Mailer: git-send-email 2.40.1 In-Reply-To: <20231122193028.GE38156@ZenIV> References: <20231122193028.GE38156@ZenIV> Precedence: bulk X-Mailing-List: linux-fsdevel@vger.kernel.org List-Id: List-Subscribe: List-Unsubscribe: MIME-Version: 1.0 Sender: Al Viro From: Jan Kara The VFS will not be locking moved directory if its parent does not change. Change reiserfs rename code to avoid touching renamed directory if its parent does not change as without locking that can corrupt the filesystem. Signed-off-by: Jan Kara Signed-off-by: Al Viro --- fs/reiserfs/namei.c | 54 ++++++++++++++++++++++++--------------------- 1 file changed, 29 insertions(+), 25 deletions(-) diff --git a/fs/reiserfs/namei.c b/fs/reiserfs/namei.c index 994d6e6995ab..5996197ba40c 100644 --- a/fs/reiserfs/namei.c +++ b/fs/reiserfs/namei.c @@ -1324,8 +1324,8 @@ static int reiserfs_rename(struct mnt_idmap *idmap, struct inode *old_inode, *new_dentry_inode; struct reiserfs_transaction_handle th; int jbegin_count; - umode_t old_inode_mode; unsigned long savelink = 1; + bool update_dir_parent = false; if (flags & ~RENAME_NOREPLACE) return -EINVAL; @@ -1375,8 +1375,7 @@ static int reiserfs_rename(struct mnt_idmap *idmap, return -ENOENT; } - old_inode_mode = old_inode->i_mode; - if (S_ISDIR(old_inode_mode)) { + if (S_ISDIR(old_inode->i_mode)) { /* * make sure that directory being renamed has correct ".." * and that its new parent directory has not too many links @@ -1389,24 +1388,28 @@ static int reiserfs_rename(struct mnt_idmap *idmap, } } - /* - * directory is renamed, its parent directory will be changed, - * so find ".." entry - */ - dot_dot_de.de_gen_number_bit_string = NULL; - retval = - reiserfs_find_entry(old_inode, "..", 2, &dot_dot_entry_path, + if (old_dir != new_dir) { + /* + * directory is renamed, its parent directory will be + * changed, so find ".." entry + */ + dot_dot_de.de_gen_number_bit_string = NULL; + retval = + reiserfs_find_entry(old_inode, "..", 2, + &dot_dot_entry_path, &dot_dot_de); - pathrelse(&dot_dot_entry_path); - if (retval != NAME_FOUND) { - reiserfs_write_unlock(old_dir->i_sb); - return -EIO; - } + pathrelse(&dot_dot_entry_path); + if (retval != NAME_FOUND) { + reiserfs_write_unlock(old_dir->i_sb); + return -EIO; + } - /* inode number of .. must equal old_dir->i_ino */ - if (dot_dot_de.de_objectid != old_dir->i_ino) { - reiserfs_write_unlock(old_dir->i_sb); - return -EIO; + /* inode number of .. must equal old_dir->i_ino */ + if (dot_dot_de.de_objectid != old_dir->i_ino) { + reiserfs_write_unlock(old_dir->i_sb); + return -EIO; + } + update_dir_parent = true; } } @@ -1486,7 +1489,7 @@ static int reiserfs_rename(struct mnt_idmap *idmap, reiserfs_prepare_for_journal(old_inode->i_sb, new_de.de_bh, 1); - if (S_ISDIR(old_inode->i_mode)) { + if (update_dir_parent) { if ((retval = search_by_entry_key(new_dir->i_sb, &dot_dot_de.de_entry_key, @@ -1534,14 +1537,14 @@ static int reiserfs_rename(struct mnt_idmap *idmap, new_de.de_bh); reiserfs_restore_prepared_buffer(old_inode->i_sb, old_de.de_bh); - if (S_ISDIR(old_inode_mode)) + if (update_dir_parent) reiserfs_restore_prepared_buffer(old_inode-> i_sb, dot_dot_de. de_bh); continue; } - if (S_ISDIR(old_inode_mode)) { + if (update_dir_parent) { if (item_moved(&dot_dot_ih, &dot_dot_entry_path) || !entry_points_to_object("..", 2, &dot_dot_de, old_dir)) { @@ -1559,7 +1562,7 @@ static int reiserfs_rename(struct mnt_idmap *idmap, } } - RFALSE(S_ISDIR(old_inode_mode) && + RFALSE(update_dir_parent && !buffer_journal_prepared(dot_dot_de.de_bh), ""); break; @@ -1592,11 +1595,12 @@ static int reiserfs_rename(struct mnt_idmap *idmap, savelink = new_dentry_inode->i_nlink; } - if (S_ISDIR(old_inode_mode)) { + if (update_dir_parent) { /* adjust ".." of renamed directory */ set_ino_in_dir_entry(&dot_dot_de, INODE_PKEY(new_dir)); journal_mark_dirty(&th, dot_dot_de.de_bh); - + } + if (S_ISDIR(old_inode->i_mode)) { /* * there (in new_dir) was no directory, so it got new link * (".." of renamed directory) From patchwork Wed Nov 22 19:36:45 2023 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: Al Viro X-Patchwork-Id: 13465417 Authentication-Results: smtp.subspace.kernel.org; dkim=pass (2048-bit key) header.d=linux.org.uk header.i=@linux.org.uk header.b="flAxY/Wb" Received: from zeniv.linux.org.uk (zeniv.linux.org.uk [IPv6:2a03:a000:7:0:5054:ff:fe1c:15ff]) by lindbergh.monkeyblade.net (Postfix) with ESMTPS id 9D1531A8; Wed, 22 Nov 2023 11:36:54 -0800 (PST) DKIM-Signature: v=1; a=rsa-sha256; q=dns/txt; c=relaxed/relaxed; d=linux.org.uk; s=zeniv-20220401; h=Sender:Content-Transfer-Encoding: MIME-Version:References:In-Reply-To:Message-Id:Date:Subject:Cc:To:From: Reply-To:Content-Type:Content-ID:Content-Description; bh=74v1Dg7yBE+4rxKY7xG9GKf0VfPvIUvY43ZeQBhdM3Q=; b=flAxY/WbiiMK3fPdqrhtjrTbxn knWsKpeOfxS/4NrPxAIcl8WOgTbfTXsv0aD4CodgCbFtRtHRh66vrZXa+5JY3Rgn7LEdSvGS1DiMO kSrZkHjLnDAcilDWdjOzrILwGHoUz/p/al5cIA2WRE9QA9jcYJgZDkgvXuukrsHhU8Db9IqZR8gma 4EbU61wQYxwy/NpdYp7IL+2NKlEDEQX/7r9fWfe7WVxxh1WvoMtG8BCEniS6YnOodUM/LCcyJ0E8i lS7BEu4tFzW7o6EBDkkcrKvhMHj4sFIRILfKgOG6AFT1RAapGtNYrUBIStplYTjITEstVZC8PnDi5 AvB/GrOw==; Received: from viro by zeniv.linux.org.uk with local (Exim 4.96 #2 (Red Hat Linux)) id 1r5t1o-001l1g-2B; Wed, 22 Nov 2023 19:36:52 +0000 From: Al Viro To: linux-fsdevel@vger.kernel.org Cc: Linus Torvalds , Mo Zou , Jan Kara , linux-kernel@vger.kernel.org Subject: [PATCH 2/9] ocfs2: Avoid touching renamed directory if parent does not change Date: Wed, 22 Nov 2023 19:36:45 +0000 Message-Id: <20231122193652.419091-2-viro@zeniv.linux.org.uk> X-Mailer: git-send-email 2.40.1 In-Reply-To: <20231122193652.419091-1-viro@zeniv.linux.org.uk> References: <20231122193028.GE38156@ZenIV> <20231122193652.419091-1-viro@zeniv.linux.org.uk> Precedence: bulk X-Mailing-List: linux-fsdevel@vger.kernel.org List-Id: List-Subscribe: List-Unsubscribe: MIME-Version: 1.0 Sender: Al Viro From: Jan Kara The VFS will not be locking moved directory if its parent does not change. Change ocfs2 rename code to avoid touching renamed directory if its parent does not change as without locking that can corrupt the filesystem. Signed-off-by: Jan Kara Signed-off-by: Al Viro --- fs/ocfs2/namei.c | 8 +++++--- 1 file changed, 5 insertions(+), 3 deletions(-) diff --git a/fs/ocfs2/namei.c b/fs/ocfs2/namei.c index 814733ba2f4b..9221a33f917b 100644 --- a/fs/ocfs2/namei.c +++ b/fs/ocfs2/namei.c @@ -1336,7 +1336,7 @@ static int ocfs2_rename(struct mnt_idmap *idmap, goto bail; } - if (S_ISDIR(old_inode->i_mode)) { + if (S_ISDIR(old_inode->i_mode) && new_dir != old_dir) { u64 old_inode_parent; update_dot_dot = 1; @@ -1353,8 +1353,7 @@ static int ocfs2_rename(struct mnt_idmap *idmap, goto bail; } - if (!new_inode && new_dir != old_dir && - new_dir->i_nlink >= ocfs2_link_max(osb)) { + if (!new_inode && new_dir->i_nlink >= ocfs2_link_max(osb)) { status = -EMLINK; goto bail; } @@ -1601,6 +1600,9 @@ static int ocfs2_rename(struct mnt_idmap *idmap, mlog_errno(status); goto bail; } + } + + if (S_ISDIR(old_inode->i_mode)) { drop_nlink(old_dir); if (new_inode) { drop_nlink(new_inode); From patchwork Wed Nov 22 19:36:46 2023 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: Al Viro X-Patchwork-Id: 13465418 Authentication-Results: smtp.subspace.kernel.org; dkim=pass (2048-bit key) header.d=linux.org.uk header.i=@linux.org.uk header.b="pG8jo4TP" Received: from zeniv.linux.org.uk (zeniv.linux.org.uk [IPv6:2a03:a000:7:0:5054:ff:fe1c:15ff]) by lindbergh.monkeyblade.net (Postfix) with ESMTPS id 9D47A1AE; Wed, 22 Nov 2023 11:36:54 -0800 (PST) DKIM-Signature: v=1; a=rsa-sha256; q=dns/txt; c=relaxed/relaxed; d=linux.org.uk; s=zeniv-20220401; h=Sender:Content-Transfer-Encoding: MIME-Version:References:In-Reply-To:Message-Id:Date:Subject:Cc:To:From: Reply-To:Content-Type:Content-ID:Content-Description; bh=rNmpOBKWoYfvVrhEukRAAHBSo1Q96GuxaczyMAbV7AU=; b=pG8jo4TPLP4OImKO1wWH4A1imn jY0RlChB5b9gcRdaSNvfTTygnHegbVfpY3EmJWx5Z6c1WUqI/Io1RU5pcNeH22yLPk54ZUe2T4eCf o7jGAJSzH3FksmBNhRWpSPA56egeiD7tUjiBFz/FIk6xNOI0PDTRmIgwzQqwuGaNueXaYKsGdKZgX LTNxN7fbMbs0T4dweEgv9ND+T8geZzkwX3NNjV1vAmzeCHrfSl3wL38z8I5eEhr/8ri0PHe5NfpVf tZwIHghZnX3G1D5CBeRIaW52QlCUgo4GEXaw/XBli2BF+5eLGwQU6PjNGOGHOtOm66D50TaHNVpnG ukkxBUZw==; Received: from viro by zeniv.linux.org.uk with local (Exim 4.96 #2 (Red Hat Linux)) id 1r5t1o-001l1i-2l; Wed, 22 Nov 2023 19:36:52 +0000 From: Al Viro To: linux-fsdevel@vger.kernel.org Cc: Linus Torvalds , Mo Zou , Jan Kara , linux-kernel@vger.kernel.org Subject: [PATCH 3/9] udf_rename(): only access the child content on cross-directory rename Date: Wed, 22 Nov 2023 19:36:46 +0000 Message-Id: <20231122193652.419091-3-viro@zeniv.linux.org.uk> X-Mailer: git-send-email 2.40.1 In-Reply-To: <20231122193652.419091-1-viro@zeniv.linux.org.uk> References: <20231122193028.GE38156@ZenIV> <20231122193652.419091-1-viro@zeniv.linux.org.uk> Precedence: bulk X-Mailing-List: linux-fsdevel@vger.kernel.org List-Id: List-Subscribe: List-Unsubscribe: MIME-Version: 1.0 Sender: Al Viro We can't really afford locking the source on same-directory rename; currently vfs_rename() tries to do that, but it will have to be changed. The logics in udf_rename() is lazy and goes looking for ".." in source even in same-directory case. It's not hard to get rid of that, leaving that behaviour only for cross-directory case; that VFS can get locks safely (and will keep doing that after the coming changes). Signed-off-by: Al Viro Reviewed-by: Jan Kara --- fs/udf/namei.c | 7 ++++++- 1 file changed, 6 insertions(+), 1 deletion(-) diff --git a/fs/udf/namei.c b/fs/udf/namei.c index 3508ac484da3..fac806a7a8d4 100644 --- a/fs/udf/namei.c +++ b/fs/udf/namei.c @@ -766,7 +766,7 @@ static int udf_rename(struct mnt_idmap *idmap, struct inode *old_dir, struct inode *old_inode = d_inode(old_dentry); struct inode *new_inode = d_inode(new_dentry); struct udf_fileident_iter oiter, niter, diriter; - bool has_diriter = false; + bool has_diriter = false, is_dir = false; int retval; struct kernel_lb_addr tloc; @@ -789,6 +789,9 @@ static int udf_rename(struct mnt_idmap *idmap, struct inode *old_dir, if (!empty_dir(new_inode)) goto out_oiter; } + is_dir = true; + } + if (is_dir && old_dir != new_dir) { retval = udf_fiiter_find_entry(old_inode, &dotdot_name, &diriter); if (retval == -ENOENT) { @@ -878,7 +881,9 @@ static int udf_rename(struct mnt_idmap *idmap, struct inode *old_dir, udf_dir_entry_len(&diriter.fi)); udf_fiiter_write_fi(&diriter, NULL); udf_fiiter_release(&diriter); + } + if (is_dir) { inode_dec_link_count(old_dir); if (new_inode) inode_dec_link_count(new_inode); From patchwork Wed Nov 22 19:36:47 2023 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: Al Viro X-Patchwork-Id: 13465414 Authentication-Results: smtp.subspace.kernel.org; dkim=pass (2048-bit key) header.d=linux.org.uk header.i=@linux.org.uk header.b="VPI2OBDP" Received: from zeniv.linux.org.uk (zeniv.linux.org.uk [IPv6:2a03:a000:7:0:5054:ff:fe1c:15ff]) by lindbergh.monkeyblade.net (Postfix) with ESMTPS id 8EB0118E; Wed, 22 Nov 2023 11:36:54 -0800 (PST) DKIM-Signature: v=1; a=rsa-sha256; q=dns/txt; c=relaxed/relaxed; d=linux.org.uk; s=zeniv-20220401; h=Sender:Content-Transfer-Encoding: MIME-Version:References:In-Reply-To:Message-Id:Date:Subject:Cc:To:From: Reply-To:Content-Type:Content-ID:Content-Description; bh=T08sUyzdKaybfEkZCNae4ocOGsjYgJj1qnspFJtuRWw=; b=VPI2OBDPeBOUVnGE8jDPzdPwIB LdSM8NnPxd857Ws5I/vrvU6Eg5/bADwz0IMxB2X9tSjNjYDE+i6ZsCxz2Py4n4oznqunCDngsEAFq foFIfg83tyMxLfEqHcMcgzNLh1yNo+5OiGshcIEUp574A8fLJKxpZ3K4y11D5PKLdfQx9ivIQmGxg xwcdf1NNJ+pMb6U7IrpOLXf1mDO4HxF7OycX466XlxKoZbAOvS6GiDonodlA32g6T6cXseWxQl4cE LFJYbxG0CxrmiXGBiYgaupA727WG9wv+oJf0YU3qE5K9PPtYhdg6YAPnWHh06MODNTSx3NeCwiUlR Z0q+NFCg==; Received: from viro by zeniv.linux.org.uk with local (Exim 4.96 #2 (Red Hat Linux)) id 1r5t1p-001l1k-05; Wed, 22 Nov 2023 19:36:53 +0000 From: Al Viro To: linux-fsdevel@vger.kernel.org Cc: Linus Torvalds , Mo Zou , Jan Kara , linux-kernel@vger.kernel.org Subject: [PATCH 4/9] ext2: Avoid reading renamed directory if parent does not change Date: Wed, 22 Nov 2023 19:36:47 +0000 Message-Id: <20231122193652.419091-4-viro@zeniv.linux.org.uk> X-Mailer: git-send-email 2.40.1 In-Reply-To: <20231122193652.419091-1-viro@zeniv.linux.org.uk> References: <20231122193028.GE38156@ZenIV> <20231122193652.419091-1-viro@zeniv.linux.org.uk> Precedence: bulk X-Mailing-List: linux-fsdevel@vger.kernel.org List-Id: List-Subscribe: List-Unsubscribe: MIME-Version: 1.0 Sender: Al Viro From: Jan Kara The VFS will not be locking moved directory if its parent does not change. Change ext2 rename code to avoid reading renamed directory if its parent does not change. Although it is currently harmless it is a bad practice to read directory contents without inode->i_rwsem. Signed-off-by: Jan Kara Signed-off-by: Al Viro --- fs/ext2/namei.c | 11 ++++++----- 1 file changed, 6 insertions(+), 5 deletions(-) diff --git a/fs/ext2/namei.c b/fs/ext2/namei.c index 65f702b1da5b..8346ab9534c1 100644 --- a/fs/ext2/namei.c +++ b/fs/ext2/namei.c @@ -325,6 +325,7 @@ static int ext2_rename (struct mnt_idmap * idmap, struct ext2_dir_entry_2 * dir_de = NULL; struct folio * old_folio; struct ext2_dir_entry_2 * old_de; + bool old_is_dir = S_ISDIR(old_inode->i_mode); int err; if (flags & ~RENAME_NOREPLACE) @@ -342,7 +343,7 @@ static int ext2_rename (struct mnt_idmap * idmap, if (IS_ERR(old_de)) return PTR_ERR(old_de); - if (S_ISDIR(old_inode->i_mode)) { + if (old_is_dir && old_dir != new_dir) { err = -EIO; dir_de = ext2_dotdot(old_inode, &dir_folio); if (!dir_de) @@ -354,7 +355,7 @@ static int ext2_rename (struct mnt_idmap * idmap, struct ext2_dir_entry_2 *new_de; err = -ENOTEMPTY; - if (dir_de && !ext2_empty_dir (new_inode)) + if (old_is_dir && !ext2_empty_dir(new_inode)) goto out_dir; new_de = ext2_find_entry(new_dir, &new_dentry->d_name, @@ -368,14 +369,14 @@ static int ext2_rename (struct mnt_idmap * idmap, if (err) goto out_dir; inode_set_ctime_current(new_inode); - if (dir_de) + if (old_is_dir) drop_nlink(new_inode); inode_dec_link_count(new_inode); } else { err = ext2_add_link(new_dentry, old_inode); if (err) goto out_dir; - if (dir_de) + if (old_is_dir) inode_inc_link_count(new_dir); } @@ -387,7 +388,7 @@ static int ext2_rename (struct mnt_idmap * idmap, mark_inode_dirty(old_inode); err = ext2_delete_entry(old_de, old_folio); - if (!err && dir_de) { + if (!err && old_is_dir) { if (old_dir != new_dir) err = ext2_set_link(old_inode, dir_de, dir_folio, new_dir, false); From patchwork Wed Nov 22 19:36:48 2023 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: Al Viro X-Patchwork-Id: 13465416 Authentication-Results: smtp.subspace.kernel.org; dkim=pass (2048-bit key) header.d=linux.org.uk header.i=@linux.org.uk header.b="LqkoXTTq" Received: from zeniv.linux.org.uk (zeniv.linux.org.uk [IPv6:2a03:a000:7:0:5054:ff:fe1c:15ff]) by lindbergh.monkeyblade.net (Postfix) with ESMTPS id 968FB19D; Wed, 22 Nov 2023 11:36:54 -0800 (PST) DKIM-Signature: v=1; a=rsa-sha256; q=dns/txt; c=relaxed/relaxed; d=linux.org.uk; s=zeniv-20220401; h=Sender:Content-Transfer-Encoding: MIME-Version:References:In-Reply-To:Message-Id:Date:Subject:Cc:To:From: Reply-To:Content-Type:Content-ID:Content-Description; bh=mSKI/kF/ejeHo+EBKx8s+JuNIL0q9sqkXoXpNGVtVqI=; b=LqkoXTTqgqMlCQAJ5QbdD+KUHW uMAKXYkMjo/D4JcVUUvgAq3BRImH56DFBksI9uMj3OT564SbAF5UmKtVQ1mACT5R1dD6tP2QTaKnn A6I82CyhEuVTfnl4RBea9t4e8xEnYF71jNq1nqcAO+HUSmtE+rtRmzgn2uOmQ6po4RJ8iv8fQmkqR hA8xrDjtFs3ygp8R/8MFE7SIuW9CH4rypmFSTy8+Bsawf+FWPXl58ktLohSYbUWkrTmtmlYxokjKW GyYYnemfn0PUD5gJISng8I7Lytq+fuY6YdS6aoEyo2LxFYoK+YRc2WEJMIEwmk3GhentERu2taLFD hihENqsg==; Received: from viro by zeniv.linux.org.uk with local (Exim 4.96 #2 (Red Hat Linux)) id 1r5t1p-001l1m-0c; Wed, 22 Nov 2023 19:36:53 +0000 From: Al Viro To: linux-fsdevel@vger.kernel.org Cc: Linus Torvalds , Mo Zou , Jan Kara , linux-kernel@vger.kernel.org Subject: [PATCH 5/9] ext4: don't access the source subdirectory content on same-directory rename Date: Wed, 22 Nov 2023 19:36:48 +0000 Message-Id: <20231122193652.419091-5-viro@zeniv.linux.org.uk> X-Mailer: git-send-email 2.40.1 In-Reply-To: <20231122193652.419091-1-viro@zeniv.linux.org.uk> References: <20231122193028.GE38156@ZenIV> <20231122193652.419091-1-viro@zeniv.linux.org.uk> Precedence: bulk X-Mailing-List: linux-fsdevel@vger.kernel.org List-Id: List-Subscribe: List-Unsubscribe: MIME-Version: 1.0 Sender: Al Viro We can't really afford locking the source on same-directory rename; currently vfs_rename() tries to do that, but it will have to be changed. The logics in ext4 is lazy and goes looking for ".." in source even in same-directory case. It's not hard to get rid of that, leaving that behaviour only for cross-directory case; that VFS can get locks safely (and will keep doing that after the coming changes). Signed-off-by: Al Viro Reviewed-by: Jan Kara --- fs/ext4/namei.c | 21 +++++++++++++-------- 1 file changed, 13 insertions(+), 8 deletions(-) diff --git a/fs/ext4/namei.c b/fs/ext4/namei.c index d252935f9c8a..467ba47a691c 100644 --- a/fs/ext4/namei.c +++ b/fs/ext4/namei.c @@ -3591,10 +3591,14 @@ struct ext4_renament { int dir_inlined; }; -static int ext4_rename_dir_prepare(handle_t *handle, struct ext4_renament *ent) +static int ext4_rename_dir_prepare(handle_t *handle, struct ext4_renament *ent, bool is_cross) { int retval; + ent->is_dir = true; + if (!is_cross) + return 0; + ent->dir_bh = ext4_get_first_dir_block(handle, ent->inode, &retval, &ent->parent_de, &ent->dir_inlined); @@ -3612,6 +3616,9 @@ static int ext4_rename_dir_finish(handle_t *handle, struct ext4_renament *ent, { int retval; + if (!ent->dir_bh) + return 0; + ent->parent_de->inode = cpu_to_le32(dir_ino); BUFFER_TRACE(ent->dir_bh, "call ext4_handle_dirty_metadata"); if (!ent->dir_inlined) { @@ -3900,7 +3907,7 @@ static int ext4_rename(struct mnt_idmap *idmap, struct inode *old_dir, if (new.dir != old.dir && EXT4_DIR_LINK_MAX(new.dir)) goto end_rename; } - retval = ext4_rename_dir_prepare(handle, &old); + retval = ext4_rename_dir_prepare(handle, &old, new.dir != old.dir); if (retval) goto end_rename; } @@ -3964,7 +3971,7 @@ static int ext4_rename(struct mnt_idmap *idmap, struct inode *old_dir, } inode_set_mtime_to_ts(old.dir, inode_set_ctime_current(old.dir)); ext4_update_dx_flag(old.dir); - if (old.dir_bh) { + if (old.is_dir) { retval = ext4_rename_dir_finish(handle, &old, new.dir->i_ino); if (retval) goto end_rename; @@ -3987,7 +3994,7 @@ static int ext4_rename(struct mnt_idmap *idmap, struct inode *old_dir, if (unlikely(retval)) goto end_rename; - if (S_ISDIR(old.inode->i_mode)) { + if (old.is_dir) { /* * We disable fast commits here that's because the * replay code is not yet capable of changing dot dot @@ -4114,14 +4121,12 @@ static int ext4_cross_rename(struct inode *old_dir, struct dentry *old_dentry, ext4_handle_sync(handle); if (S_ISDIR(old.inode->i_mode)) { - old.is_dir = true; - retval = ext4_rename_dir_prepare(handle, &old); + retval = ext4_rename_dir_prepare(handle, &old, new.dir != old.dir); if (retval) goto end_rename; } if (S_ISDIR(new.inode->i_mode)) { - new.is_dir = true; - retval = ext4_rename_dir_prepare(handle, &new); + retval = ext4_rename_dir_prepare(handle, &new, new.dir != old.dir); if (retval) goto end_rename; } From patchwork Wed Nov 22 19:36:49 2023 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: Al Viro X-Patchwork-Id: 13465419 Authentication-Results: smtp.subspace.kernel.org; dkim=pass (2048-bit key) header.d=linux.org.uk header.i=@linux.org.uk header.b="j/zHIGdb" Received: from zeniv.linux.org.uk (zeniv.linux.org.uk [IPv6:2a03:a000:7:0:5054:ff:fe1c:15ff]) by lindbergh.monkeyblade.net (Postfix) with ESMTPS id 9DE2B1B2; Wed, 22 Nov 2023 11:36:54 -0800 (PST) DKIM-Signature: v=1; a=rsa-sha256; q=dns/txt; c=relaxed/relaxed; d=linux.org.uk; s=zeniv-20220401; h=Sender:Content-Transfer-Encoding: MIME-Version:References:In-Reply-To:Message-Id:Date:Subject:Cc:To:From: Reply-To:Content-Type:Content-ID:Content-Description; bh=1kXdQTzP9QbMwip9HhQkODeDCTy6uPz4pJDbGLrfwH4=; b=j/zHIGdbj57GVSf3DgjFtbdzKE YXPaab07VjQSO1hmibpG7Yx3cWdhR6Gt7FJ3w44L7KwZ+vQ16Xm2JYzQ47NgaIioMCP6/HINChePB f9BnMFXX3+IZzeLoCwZmE+JGw6KJqao1AGBWgNv7jG1nZWOI355W9J63abY9m5dxqknDvV1hLqZf6 S//r8xOmWheYW872BRLirbQJW8c8LFS7D2t+uNB2TradI9dAqRwrQoxFgxiblODgGndnFa92Jvt3v zExswAs4CxD3Z7cXEkzeINfSkXTYumauvHRuO8aJtGzPAuZp7m7PV8jpHXBNK6+bSAf0lNVZTgzEH P/mGlS3w==; Received: from viro by zeniv.linux.org.uk with local (Exim 4.96 #2 (Red Hat Linux)) id 1r5t1p-001l1o-0z; Wed, 22 Nov 2023 19:36:53 +0000 From: Al Viro To: linux-fsdevel@vger.kernel.org Cc: Linus Torvalds , Mo Zou , Jan Kara , linux-kernel@vger.kernel.org Subject: [PATCH 6/9] f2fs: Avoid reading renamed directory if parent does not change Date: Wed, 22 Nov 2023 19:36:49 +0000 Message-Id: <20231122193652.419091-6-viro@zeniv.linux.org.uk> X-Mailer: git-send-email 2.40.1 In-Reply-To: <20231122193652.419091-1-viro@zeniv.linux.org.uk> References: <20231122193028.GE38156@ZenIV> <20231122193652.419091-1-viro@zeniv.linux.org.uk> Precedence: bulk X-Mailing-List: linux-fsdevel@vger.kernel.org List-Id: List-Subscribe: List-Unsubscribe: MIME-Version: 1.0 Sender: Al Viro From: Jan Kara The VFS will not be locking moved directory if its parent does not change. Change f2fs rename code to avoid reading renamed directory if its parent does not change. Having it uninlined while we are reading it would cause trouble and we won't be able to rely upon ->i_rwsem on the directory being renamed in cases that do not alter its parent. Signed-off-by: Jan Kara Signed-off-by: Al Viro --- fs/f2fs/namei.c | 15 ++++++++------- 1 file changed, 8 insertions(+), 7 deletions(-) diff --git a/fs/f2fs/namei.c b/fs/f2fs/namei.c index d0053b0284d8..fdc97df6bb85 100644 --- a/fs/f2fs/namei.c +++ b/fs/f2fs/namei.c @@ -963,6 +963,7 @@ static int f2fs_rename(struct mnt_idmap *idmap, struct inode *old_dir, struct f2fs_dir_entry *old_dir_entry = NULL; struct f2fs_dir_entry *old_entry; struct f2fs_dir_entry *new_entry; + bool old_is_dir = S_ISDIR(old_inode->i_mode); int err; if (unlikely(f2fs_cp_error(sbi))) @@ -1017,7 +1018,7 @@ static int f2fs_rename(struct mnt_idmap *idmap, struct inode *old_dir, goto out; } - if (S_ISDIR(old_inode->i_mode)) { + if (old_is_dir && old_dir != new_dir) { old_dir_entry = f2fs_parent_dir(old_inode, &old_dir_page); if (!old_dir_entry) { if (IS_ERR(old_dir_page)) @@ -1029,7 +1030,7 @@ static int f2fs_rename(struct mnt_idmap *idmap, struct inode *old_dir, if (new_inode) { err = -ENOTEMPTY; - if (old_dir_entry && !f2fs_empty_dir(new_inode)) + if (old_is_dir && !f2fs_empty_dir(new_inode)) goto out_dir; err = -ENOENT; @@ -1054,7 +1055,7 @@ static int f2fs_rename(struct mnt_idmap *idmap, struct inode *old_dir, inode_set_ctime_current(new_inode); f2fs_down_write(&F2FS_I(new_inode)->i_sem); - if (old_dir_entry) + if (old_is_dir) f2fs_i_links_write(new_inode, false); f2fs_i_links_write(new_inode, false); f2fs_up_write(&F2FS_I(new_inode)->i_sem); @@ -1074,12 +1075,12 @@ static int f2fs_rename(struct mnt_idmap *idmap, struct inode *old_dir, goto out_dir; } - if (old_dir_entry) + if (old_is_dir) f2fs_i_links_write(new_dir, true); } f2fs_down_write(&F2FS_I(old_inode)->i_sem); - if (!old_dir_entry || whiteout) + if (!old_is_dir || whiteout) file_lost_pino(old_inode); else /* adjust dir's i_pino to pass fsck check */ @@ -1105,8 +1106,8 @@ static int f2fs_rename(struct mnt_idmap *idmap, struct inode *old_dir, iput(whiteout); } - if (old_dir_entry) { - if (old_dir != new_dir && !whiteout) + if (old_is_dir) { + if (old_dir_entry && !whiteout) f2fs_set_link(old_inode, old_dir_entry, old_dir_page, new_dir); else From patchwork Wed Nov 22 19:36:50 2023 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: Al Viro X-Patchwork-Id: 13465421 Authentication-Results: smtp.subspace.kernel.org; dkim=pass (2048-bit key) header.d=linux.org.uk header.i=@linux.org.uk header.b="DQlLrAJ9" Received: from zeniv.linux.org.uk (zeniv.linux.org.uk [IPv6:2a03:a000:7:0:5054:ff:fe1c:15ff]) by lindbergh.monkeyblade.net (Postfix) with ESMTPS id D833A1B5; Wed, 22 Nov 2023 11:36:54 -0800 (PST) DKIM-Signature: v=1; a=rsa-sha256; q=dns/txt; c=relaxed/relaxed; d=linux.org.uk; s=zeniv-20220401; h=Sender:Content-Transfer-Encoding: MIME-Version:References:In-Reply-To:Message-Id:Date:Subject:Cc:To:From: Reply-To:Content-Type:Content-ID:Content-Description; bh=2Ih+XGz67G/g1oREsd3pNbBpUO9/7/Wqei4gcLL9Zq8=; b=DQlLrAJ9HgsRqZMo+pE35Io3lM mJkanKmP4JmFqOriWCBDgTPVYaaAls7n31yHBIvgRdjVli8qpL0MEmXOnP8GJuoOr3hr8A/edD7KW udJAPMID8KvgVzhIu28qvl/mr9kso2tmmUcjFstT8M0wHjBuZemn+1H+qvYu8xnzxN9vyohuHHuVe icfwVk+PWnDpHHRDtNN4+8DCkwNCbWldU1pRzlwX/HdORE5tg7z3XFag7uxiizEYBKZIDXq0U5BCq z9FYaqm409dJ5bBvmUMF0NbZkn6ilSasW6gHWbG/5oONxHgnoFj9yMNkDeRhNqjqHBRLcGRPTtfsw CIH7EbFw==; Received: from viro by zeniv.linux.org.uk with local (Exim 4.96 #2 (Red Hat Linux)) id 1r5t1p-001l20-1M; Wed, 22 Nov 2023 19:36:53 +0000 From: Al Viro To: linux-fsdevel@vger.kernel.org Cc: Linus Torvalds , Mo Zou , Jan Kara , linux-kernel@vger.kernel.org Subject: [PATCH 7/9] rename(): fix the locking of subdirectories Date: Wed, 22 Nov 2023 19:36:50 +0000 Message-Id: <20231122193652.419091-7-viro@zeniv.linux.org.uk> X-Mailer: git-send-email 2.40.1 In-Reply-To: <20231122193652.419091-1-viro@zeniv.linux.org.uk> References: <20231122193028.GE38156@ZenIV> <20231122193652.419091-1-viro@zeniv.linux.org.uk> Precedence: bulk X-Mailing-List: linux-fsdevel@vger.kernel.org List-Id: List-Subscribe: List-Unsubscribe: MIME-Version: 1.0 Sender: Al Viro We should never lock two subdirectories without having taken ->s_vfs_rename_mutex; inode pointer order or not, the "order" proposed in 28eceeda130f "fs: Lock moved directories" is not transitive, with the usual consequences. The rationale for locking renamed subdirectory in all cases was the possibility of race between rename modifying .. in a subdirectory to reflect the new parent and another thread modifying the same subdirectory. For a lot of filesystems that's not a problem, but for some it can lead to trouble (e.g. the case when short directory contents is kept in the inode, but creating a file in it might push it across the size limit and copy its contents into separate data block(s)). However, we need that only in case when the parent does change - otherwise ->rename() doesn't need to do anything with .. entry in the first place. Some instances are lazy and do a tautological update anyway, but it's really not hard to avoid. Amended locking rules for rename(): find the parent(s) of source and target if source and target have the same parent lock the common parent else lock ->s_vfs_rename_mutex lock both parents, in ancestor-first order; if neither is an ancestor of another, lock the parent of source first. find the source and target. if source and target have the same parent if operation is an overwriting rename of a subdirectory lock the target subdirectory else if source is a subdirectory lock the source if target is a subdirectory lock the target lock non-directories involved, in inode pointer order if both source and target are such. That way we are guaranteed that parents are locked (for obvious reasons), that any renamed non-directory is locked (nfsd relies upon that), that any victim is locked (emptiness check needs that, among other things) and subdirectory that changes parent is locked (needed to protect the update of .. entries). We are also guaranteed that any operation locking more than one directory either takes ->s_vfs_rename_mutex or locks a parent followed by its child. Cc: stable@vger.kernel.org Fixes: 28eceeda130f "fs: Lock moved directories" Signed-off-by: Al Viro Reviewed-by: Jan Kara --- .../filesystems/directory-locking.rst | 29 ++++----- Documentation/filesystems/locking.rst | 5 +- Documentation/filesystems/porting.rst | 18 ++++++ fs/namei.c | 60 ++++++++++++------- 4 files changed, 74 insertions(+), 38 deletions(-) diff --git a/Documentation/filesystems/directory-locking.rst b/Documentation/filesystems/directory-locking.rst index dccd61c7c5c3..193c22687851 100644 --- a/Documentation/filesystems/directory-locking.rst +++ b/Documentation/filesystems/directory-locking.rst @@ -22,13 +22,16 @@ exclusive. 3) object removal. Locking rules: caller locks parent, finds victim, locks victim and calls the method. Locks are exclusive. -4) rename() that is _not_ cross-directory. Locking rules: caller locks the -parent and finds source and target. We lock both (provided they exist). If we -need to lock two inodes of different type (dir vs non-dir), we lock directory -first. If we need to lock two inodes of the same type, lock them in inode -pointer order. Then call the method. All locks are exclusive. -NB: we might get away with locking the source (and target in exchange -case) shared. +4) rename() that is _not_ cross-directory. Locking rules: caller locks +the parent and finds source and target. Then we decide which of the +source and target need to be locked. Source needs to be locked if it's a +non-directory; target - if it's a non-directory or about to be removed. +Take the locks that need to be taken, in inode pointer order if need +to take both (that can happen only when both source and target are +non-directories - the source because it wouldn't be locked otherwise +and the target because mixing directory and non-directory is allowed +only with RENAME_EXCHANGE, and that won't be removing the target). +After the locks had been taken, call the method. All locks are exclusive. 5) link creation. Locking rules: @@ -44,20 +47,17 @@ rules: * lock the filesystem * lock parents in "ancestors first" order. If one is not ancestor of - the other, lock them in inode pointer order. + the other, lock the parent of source first. * find source and target. * if old parent is equal to or is a descendent of target fail with -ENOTEMPTY * if new parent is equal to or is a descendent of source fail with -ELOOP - * Lock both the source and the target provided they exist. If we - need to lock two inodes of different type (dir vs non-dir), we lock - the directory first. If we need to lock two inodes of the same type, - lock them in inode pointer order. + * Lock subdirectories involved (source before target). + * Lock non-directories involved, in inode pointer order. * call the method. -All ->i_rwsem are taken exclusive. Again, we might get away with locking -the source (and target in exchange case) shared. +All ->i_rwsem are taken exclusive. The rules above obviously guarantee that all directories that are going to be read, modified or removed by method will be locked by caller. @@ -67,6 +67,7 @@ If no directory is its own ancestor, the scheme above is deadlock-free. Proof: +[XXX: will be updated once we are done massaging the lock_rename()] First of all, at any moment we have a linear ordering of the objects - A < B iff (A is an ancestor of B) or (B is not an ancestor of A and ptr(A) < ptr(B)). diff --git a/Documentation/filesystems/locking.rst b/Documentation/filesystems/locking.rst index 7be2900806c8..bd12f2f850ad 100644 --- a/Documentation/filesystems/locking.rst +++ b/Documentation/filesystems/locking.rst @@ -101,7 +101,7 @@ symlink: exclusive mkdir: exclusive unlink: exclusive (both) rmdir: exclusive (both)(see below) -rename: exclusive (all) (see below) +rename: exclusive (both parents, some children) (see below) readlink: no get_link: no setattr: exclusive @@ -123,6 +123,9 @@ get_offset_ctx no Additionally, ->rmdir(), ->unlink() and ->rename() have ->i_rwsem exclusive on victim. cross-directory ->rename() has (per-superblock) ->s_vfs_rename_sem. + ->unlink() and ->rename() have ->i_rwsem exclusive on all non-directories + involved. + ->rename() has ->i_rwsem exclusive on any subdirectory that changes parent. See Documentation/filesystems/directory-locking.rst for more detailed discussion of the locking scheme for directory operations. diff --git a/Documentation/filesystems/porting.rst b/Documentation/filesystems/porting.rst index 878e72b2f8b7..9100969e7de6 100644 --- a/Documentation/filesystems/porting.rst +++ b/Documentation/filesystems/porting.rst @@ -1061,3 +1061,21 @@ export_operations ->encode_fh() no longer has a default implementation to encode FILEID_INO32_GEN* file handles. Filesystems that used the default implementation may use the generic helper generic_encode_ino32_fh() explicitly. + +--- + +**mandatory** + +If ->rename() update of .. on cross-directory move needs an exclusion with +directory modifications, do *not* lock the subdirectory in question in your +->rename() - it's done by the caller now [that item should've been added in +28eceeda130f "fs: Lock moved directories"]. + +--- + +**mandatory** + +On same-directory ->rename() the (tautological) update of .. is not protected +by any locks; just don't do it if the old parent is the same as the new one. +We really can't lock two subdirectories in same-directory rename - not without +deadlocks. diff --git a/fs/namei.c b/fs/namei.c index 71c13b2990b4..29bafbdb44ca 100644 --- a/fs/namei.c +++ b/fs/namei.c @@ -3021,20 +3021,14 @@ static struct dentry *lock_two_directories(struct dentry *p1, struct dentry *p2) p = d_ancestor(p2, p1); if (p) { inode_lock_nested(p2->d_inode, I_MUTEX_PARENT); - inode_lock_nested(p1->d_inode, I_MUTEX_CHILD); + inode_lock_nested(p1->d_inode, I_MUTEX_PARENT2); return p; } p = d_ancestor(p1, p2); - if (p) { - inode_lock_nested(p1->d_inode, I_MUTEX_PARENT); - inode_lock_nested(p2->d_inode, I_MUTEX_CHILD); - return p; - } - - lock_two_inodes(p1->d_inode, p2->d_inode, - I_MUTEX_PARENT, I_MUTEX_PARENT2); - return NULL; + inode_lock_nested(p1->d_inode, I_MUTEX_PARENT); + inode_lock_nested(p2->d_inode, I_MUTEX_PARENT2); + return p; } /* @@ -4716,11 +4710,12 @@ SYSCALL_DEFINE2(link, const char __user *, oldname, const char __user *, newname * * a) we can get into loop creation. * b) race potential - two innocent renames can create a loop together. - * That's where 4.4 screws up. Current fix: serialization on + * That's where 4.4BSD screws up. Current fix: serialization on * sb->s_vfs_rename_mutex. We might be more accurate, but that's another * story. - * c) we have to lock _four_ objects - parents and victim (if it exists), - * and source. + * c) we may have to lock up to _four_ objects - parents and victim (if it exists), + * and source (if it's a non-directory or a subdirectory that moves to + * different parent). * And that - after we got ->i_mutex on parents (until then we don't know * whether the target exists). Solution: try to be smart with locking * order for inodes. We rely on the fact that tree topology may change @@ -4752,6 +4747,7 @@ int vfs_rename(struct renamedata *rd) bool new_is_dir = false; unsigned max_links = new_dir->i_sb->s_max_links; struct name_snapshot old_name; + bool lock_old_subdir, lock_new_subdir; if (source == target) return 0; @@ -4805,15 +4801,32 @@ int vfs_rename(struct renamedata *rd) take_dentry_name_snapshot(&old_name, old_dentry); dget(new_dentry); /* - * Lock all moved children. Moved directories may need to change parent - * pointer so they need the lock to prevent against concurrent - * directory changes moving parent pointer. For regular files we've - * historically always done this. The lockdep locking subclasses are - * somewhat arbitrary but RENAME_EXCHANGE in particular can swap - * regular files and directories so it's difficult to tell which - * subclasses to use. + * Lock children. + * The source subdirectory needs to be locked on cross-directory + * rename or cross-directory exchange since its parent changes. + * The target subdirectory needs to be locked on cross-directory + * exchange due to parent change and on any rename due to becoming + * a victim. + * Non-directories need locking in all cases (for NFS reasons); + * they get locked after any subdirectories (in inode address order). + * + * NOTE: WE ONLY LOCK UNRELATED DIRECTORIES IN CROSS-DIRECTORY CASE. + * NEVER, EVER DO THAT WITHOUT ->s_vfs_rename_mutex. */ - lock_two_inodes(source, target, I_MUTEX_NORMAL, I_MUTEX_NONDIR2); + lock_old_subdir = new_dir != old_dir; + lock_new_subdir = new_dir != old_dir || !(flags & RENAME_EXCHANGE); + if (is_dir) { + if (lock_old_subdir) + inode_lock_nested(source, I_MUTEX_CHILD); + if (target && (!new_is_dir || lock_new_subdir)) + inode_lock(target); + } else if (new_is_dir) { + if (lock_new_subdir) + inode_lock_nested(target, I_MUTEX_CHILD); + inode_lock(source); + } else { + lock_two_nondirectories(source, target); + } error = -EPERM; if (IS_SWAPFILE(source) || (target && IS_SWAPFILE(target))) @@ -4861,8 +4874,9 @@ int vfs_rename(struct renamedata *rd) d_exchange(old_dentry, new_dentry); } out: - inode_unlock(source); - if (target) + if (!is_dir || lock_old_subdir) + inode_unlock(source); + if (target && (!new_is_dir || lock_new_subdir)) inode_unlock(target); dput(new_dentry); if (!error) { From patchwork Wed Nov 22 19:36:51 2023 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: Al Viro X-Patchwork-Id: 13465420 Authentication-Results: smtp.subspace.kernel.org; dkim=pass (2048-bit key) header.d=linux.org.uk header.i=@linux.org.uk header.b="C35Pzkvi" Received: from zeniv.linux.org.uk (zeniv.linux.org.uk [IPv6:2a03:a000:7:0:5054:ff:fe1c:15ff]) by lindbergh.monkeyblade.net (Postfix) with ESMTPS id E18AC1B6; Wed, 22 Nov 2023 11:36:54 -0800 (PST) DKIM-Signature: v=1; a=rsa-sha256; q=dns/txt; c=relaxed/relaxed; d=linux.org.uk; s=zeniv-20220401; h=Sender:Content-Transfer-Encoding: MIME-Version:References:In-Reply-To:Message-Id:Date:Subject:Cc:To:From: Reply-To:Content-Type:Content-ID:Content-Description; bh=8QOQpop95kL18saXqv/BTfGcqZ0IcnSd0yUZGFIOYns=; b=C35Pzkvi4HIowDl9UXBA5b+YHW C+sNIcDjX6bYZayTNn5EehZyXWftX+Gl1rMAOvyYClQuWLiRaCqo4cyPYPCyBDMF38WP+WSiKxHld YvhpNg59pNQw4rbRLRmWpMhhBimau0x/2JXNpHJPqFxrROBxvFHlND3OPyIZ9tlJY+BZeyg8CJ8aT CuLvxELVcWkzvrB7JqMN9kpgNGjzhSXUYXpr5H+5pc+OzpUZjk25yJ4BPztH10ohblVsYkrP6Ln3L imJVt9DatNbrjJDgYaZX6l9DSst+EfsoM7R8ssMo9UxQ3f4z0IgCURffIHG98QPJjgRlMETAFscwb cNQkZckg==; Received: from viro by zeniv.linux.org.uk with local (Exim 4.96 #2 (Red Hat Linux)) id 1r5t1p-001l24-1t; Wed, 22 Nov 2023 19:36:53 +0000 From: Al Viro To: linux-fsdevel@vger.kernel.org Cc: Linus Torvalds , Mo Zou , Jan Kara , linux-kernel@vger.kernel.org Subject: [PATCH 8/9] kill lock_two_inodes() Date: Wed, 22 Nov 2023 19:36:51 +0000 Message-Id: <20231122193652.419091-8-viro@zeniv.linux.org.uk> X-Mailer: git-send-email 2.40.1 In-Reply-To: <20231122193652.419091-1-viro@zeniv.linux.org.uk> References: <20231122193028.GE38156@ZenIV> <20231122193652.419091-1-viro@zeniv.linux.org.uk> Precedence: bulk X-Mailing-List: linux-fsdevel@vger.kernel.org List-Id: List-Subscribe: List-Unsubscribe: MIME-Version: 1.0 Sender: Al Viro There's only one caller left (lock_two_nondirectories()), and it needs less complexity. Fold lock_two_inodes() in there and simplify. Signed-off-by: Al Viro Reviewed-by: Jan Kara --- fs/inode.c | 49 ++++++------------------------------------------- fs/internal.h | 2 -- 2 files changed, 6 insertions(+), 45 deletions(-) diff --git a/fs/inode.c b/fs/inode.c index edcd8a61975f..453d5be1a014 100644 --- a/fs/inode.c +++ b/fs/inode.c @@ -1087,48 +1087,6 @@ void discard_new_inode(struct inode *inode) } EXPORT_SYMBOL(discard_new_inode); -/** - * lock_two_inodes - lock two inodes (may be regular files but also dirs) - * - * Lock any non-NULL argument. The caller must make sure that if he is passing - * in two directories, one is not ancestor of the other. Zero, one or two - * objects may be locked by this function. - * - * @inode1: first inode to lock - * @inode2: second inode to lock - * @subclass1: inode lock subclass for the first lock obtained - * @subclass2: inode lock subclass for the second lock obtained - */ -void lock_two_inodes(struct inode *inode1, struct inode *inode2, - unsigned subclass1, unsigned subclass2) -{ - if (!inode1 || !inode2) { - /* - * Make sure @subclass1 will be used for the acquired lock. - * This is not strictly necessary (no current caller cares) but - * let's keep things consistent. - */ - if (!inode1) - swap(inode1, inode2); - goto lock; - } - - /* - * If one object is directory and the other is not, we must make sure - * to lock directory first as the other object may be its child. - */ - if (S_ISDIR(inode2->i_mode) == S_ISDIR(inode1->i_mode)) { - if (inode1 > inode2) - swap(inode1, inode2); - } else if (!S_ISDIR(inode1->i_mode)) - swap(inode1, inode2); -lock: - if (inode1) - inode_lock_nested(inode1, subclass1); - if (inode2 && inode2 != inode1) - inode_lock_nested(inode2, subclass2); -} - /** * lock_two_nondirectories - take two i_mutexes on non-directory objects * @@ -1144,7 +1102,12 @@ void lock_two_nondirectories(struct inode *inode1, struct inode *inode2) WARN_ON_ONCE(S_ISDIR(inode1->i_mode)); if (inode2) WARN_ON_ONCE(S_ISDIR(inode2->i_mode)); - lock_two_inodes(inode1, inode2, I_MUTEX_NORMAL, I_MUTEX_NONDIR2); + if (inode1 > inode2) + swap(inode1, inode2); + if (inode1) + inode_lock(inode1); + if (inode2 && inode2 != inode1) + inode_lock_nested(inode2, I_MUTEX_NONDIR2); } EXPORT_SYMBOL(lock_two_nondirectories); diff --git a/fs/internal.h b/fs/internal.h index 58e43341aebf..de67b02226e5 100644 --- a/fs/internal.h +++ b/fs/internal.h @@ -196,8 +196,6 @@ extern long prune_icache_sb(struct super_block *sb, struct shrink_control *sc); int dentry_needs_remove_privs(struct mnt_idmap *, struct dentry *dentry); bool in_group_or_capable(struct mnt_idmap *idmap, const struct inode *inode, vfsgid_t vfsgid); -void lock_two_inodes(struct inode *inode1, struct inode *inode2, - unsigned subclass1, unsigned subclass2); /* * fs-writeback.c From patchwork Wed Nov 22 19:36:52 2023 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: Al Viro X-Patchwork-Id: 13465422 Authentication-Results: smtp.subspace.kernel.org; dkim=pass (2048-bit key) header.d=linux.org.uk header.i=@linux.org.uk header.b="GNxkxzIO" Received: from zeniv.linux.org.uk (zeniv.linux.org.uk [IPv6:2a03:a000:7:0:5054:ff:fe1c:15ff]) by lindbergh.monkeyblade.net (Postfix) with ESMTPS id 2099D1B9; Wed, 22 Nov 2023 11:36:55 -0800 (PST) DKIM-Signature: v=1; a=rsa-sha256; q=dns/txt; c=relaxed/relaxed; d=linux.org.uk; s=zeniv-20220401; h=Sender:Content-Transfer-Encoding: MIME-Version:References:In-Reply-To:Message-Id:Date:Subject:Cc:To:From: Reply-To:Content-Type:Content-ID:Content-Description; bh=YMKZBxnotjPbPEkf1tqWwQKMdfRNNI8K9vfV89aUGzY=; b=GNxkxzIOiCyt14ZAVuwPyxjL4S EELt5mxZuv+cHX6WR/4a+delBpo3T2IoiThPtRefL/KTBH+/zGTEIJF1sYHXXso9mkZncrAwTTNQY T8sedrTJiTzPcNvkCb8B0KsrdWkSDqPX2Cdhd3flfjlWcFPM+tvHtYD5ZnOHsLZ+ShNlR23P9TTdR mlgXUFLRYnaux33geeRvmLriDgwEMsJBu47HKEt32k0/I2Ygd0ti9sGvSSWisSX2c6Dg6GDmFTU78 mXwdRlJEZ9NpS0/UkDjXX+Io5b7ijej+OFTDF5oyO6iaMW7JyXWb0eBykpGWudfn76ElCzVWWkBqj 19DjdvNg==; Received: from viro by zeniv.linux.org.uk with local (Exim 4.96 #2 (Red Hat Linux)) id 1r5t1p-001l28-2C; Wed, 22 Nov 2023 19:36:53 +0000 From: Al Viro To: linux-fsdevel@vger.kernel.org Cc: Linus Torvalds , Mo Zou , Jan Kara , linux-kernel@vger.kernel.org Subject: [PATCH 9/9] rename(): avoid a deadlock in the case of parents having no common ancestor Date: Wed, 22 Nov 2023 19:36:52 +0000 Message-Id: <20231122193652.419091-9-viro@zeniv.linux.org.uk> X-Mailer: git-send-email 2.40.1 In-Reply-To: <20231122193652.419091-1-viro@zeniv.linux.org.uk> References: <20231122193028.GE38156@ZenIV> <20231122193652.419091-1-viro@zeniv.linux.org.uk> Precedence: bulk X-Mailing-List: linux-fsdevel@vger.kernel.org List-Id: List-Subscribe: List-Unsubscribe: MIME-Version: 1.0 Sender: Al Viro ... and fix the directory locking documentation and proof of correctness. Holding ->s_vfs_rename_mutex *almost* prevents ->d_parent changes; the case where we really don't want it is splicing the root of disconnected tree to somewhere. In other words, ->s_vfs_rename_mutex is sufficient to stabilize "X is an ancestor of Y" only if X and Y are already in the same tree. Otherwise it can go from false to true, and one can construct a deadlock on that. Make lock_two_directories() report an error in such case and update the callers of lock_rename()/lock_rename_child() to handle such errors. And yes, such conditions are not impossible to create ;-/ Signed-off-by: Al Viro Reviewed-by: Jan Kara Acked-by: Amir Goldstein --- .../filesystems/directory-locking.rst | 302 ++++++++++++------ Documentation/filesystems/porting.rst | 9 + fs/cachefiles/namei.c | 2 + fs/ecryptfs/inode.c | 2 + fs/namei.c | 37 ++- fs/nfsd/vfs.c | 4 + fs/overlayfs/copy_up.c | 9 +- fs/overlayfs/dir.c | 4 + fs/overlayfs/super.c | 6 +- fs/overlayfs/util.c | 7 +- fs/smb/server/vfs.c | 5 + 11 files changed, 276 insertions(+), 111 deletions(-) diff --git a/Documentation/filesystems/directory-locking.rst b/Documentation/filesystems/directory-locking.rst index 193c22687851..9bf497539eb0 100644 --- a/Documentation/filesystems/directory-locking.rst +++ b/Documentation/filesystems/directory-locking.rst @@ -11,121 +11,230 @@ When taking the i_rwsem on multiple non-directory objects, we always acquire the locks in order by increasing address. We'll call that "inode pointer" order in the following. -For our purposes all operations fall in 5 classes: -1) read access. Locking rules: caller locks directory we are accessing. -The lock is taken shared. + Primitives. -2) object creation. Locking rules: same as above, but the lock is taken -exclusive. +For our purposes all operations fall in 6 classes: -3) object removal. Locking rules: caller locks parent, finds victim, -locks victim and calls the method. Locks are exclusive. +1) read access. Locking rules: caller locks the directory we are +accessing. The lock is taken shared. -4) rename() that is _not_ cross-directory. Locking rules: caller locks -the parent and finds source and target. Then we decide which of the -source and target need to be locked. Source needs to be locked if it's a -non-directory; target - if it's a non-directory or about to be removed. -Take the locks that need to be taken, in inode pointer order if need -to take both (that can happen only when both source and target are -non-directories - the source because it wouldn't be locked otherwise -and the target because mixing directory and non-directory is allowed -only with RENAME_EXCHANGE, and that won't be removing the target). -After the locks had been taken, call the method. All locks are exclusive. +2) object creation. Locking rules: same as above, but directory lock +is taken exclusive. -5) link creation. Locking rules: +3) object removal. Locking rules: caller locks the parent, finds the +victim, locks the victim and calls the method. Locks are exclusive. - * lock parent - * check that source is not a directory - * lock source +4) link creation. Locking rules: + * lock the parent + * check that the source is not a directory + * lock the source * call the method. +All locks are exclusive. +5) rename() that is _not_ cross-directory. Locking rules: + * lock the parent + * find the source and target. + * decide which of the source and target need to be locked. +The source needs to be locked if it's a non-directory, target - if it's +a non-directory or about to be removed. Take the locks that need to be +taken, in inode pointer order if need to take both (that can happen only +when both source and target are non-directories - the source because +it wouldn't need to be locked otherwise and the target because mixing +directory and non-directory is allowed only with RENAME_EXCHANGE, and +that won't be removing the target). + * call the method. All locks are exclusive. 6) cross-directory rename. The trickiest in the whole bunch. Locking rules: - * lock the filesystem - * lock parents in "ancestors first" order. If one is not ancestor of - the other, lock the parent of source first. - * find source and target. - * if old parent is equal to or is a descendent of target - fail with -ENOTEMPTY - * if new parent is equal to or is a descendent of source - fail with -ELOOP - * Lock subdirectories involved (source before target). - * Lock non-directories involved, in inode pointer order. + * if the parents don't have a common ancestor, fail the operation. + * lock the parents in "ancestors first" order. If neither is an +ancestor of the other, lock the parent of source first. + * find the source and target. + * verify that the source is not a descendent of the target and +target is not a descendent of source; fail the operation otherwise. + * lock the subdirectories involved (source before target). + * lock the non-directories involved, in inode pointer order. * call the method. - All ->i_rwsem are taken exclusive. -The rules above obviously guarantee that all directories that are going to be -read, modified or removed by method will be locked by caller. - +The rules above obviously guarantee that all directories that are going +to be read, modified or removed by method will be locked by the caller. + + + Splicing. + +There is one more thing to consider - splicing. It's not an operation +in its own right; it may happen as part of lookup. We speak of the +operations on directory trees, but we obviously do not have the full +picture of those - especially for network filesystems. What we have +is a bunch of subtrees visible in dcache and locking happens on those. +Trees grow as we do operations; memory pressure prunes them. Normally +that's not a problem, but there is a nasty twist - what should we do +when one growing tree reaches the root of another? That can happen in +several scenarios, starting from "somebody mounted two nested subtrees +from the same NFS4 server and doing lookups in one of them has reached +the root of another"; there's also open-by-fhandle stuff, and there's a +possibility that directory we see in one place gets moved by the server +to another and we run into it when we do a lookup. + +For a lot of reasons we want to have the same directory present in +dcache only once. Multiple aliases are not allowed. So when lookup +runs into a subdirectory that already has an alias, something needs to +be done with dcache trees. Lookup is already holding the parent locked. +If alias is a root of separate tree, it gets attached to the directory +we are doing a lookup in, under the name we'd been looking for. +If the alias is already a child of the directory we are looking in, +it changes name to the one we'd been looking for. No extra locking is +involved in these two cases. However, if it's a child of some other +directory, the things get trickier. First of all, we verify that it +is *not* an ancestor of our directory and fail the lookup if it is. +Then we try to lock the filesystem and the current parent of the alias. +If either trylock fails, we fail the lookup. If trylocks succeed, +we detach the alias from its current parent and attach to our directory, +under the name we are looking for. + +Note that splicing does *not* involve any modification of the filesystem; +all we change is the view in dcache. Moreover, holding a directory +locked exclusive prevents such changes involving its children and holding +the filesystem lock prevents any changes of tree topology, other than +having a root of one tree becoming a child of directory in another. In +particular, if two dentries have been found to have a common ancestor +after taking the filesystem lock, their relationship will remain unchanged +until the lock is dropped. So from the directory operations' point of +view splicing is almost irrelevant - the only place where it matters is +one step in cross-directory renames; we need to be careful when checking +if parents have a common ancestor. + + + Multiple-filesystem stuff. + +For some filesystems a method can involve a directory operation on +another filesystem; it may be ecryptfs doing operation in the underlying +filesystem, overlayfs doing something to the layers, network filesystem +using a local one as a cache, etc. In all such cases the operations +on other filesystems must follow the same locking rules. Moreover, "a +directory operation on this filesystem might involve directory operations +on that filesystem" should be an asymmetric relation (or, if you will, +it should be possible to rank the filesystems so that directory operation +on a filesystem could trigger directory operations only on higher-ranked +ones - in these terms overlayfs ranks lower than its layers, network +filesystem ranks lower than whatever it caches on, etc.) + + + Deadlock avoidance. If no directory is its own ancestor, the scheme above is deadlock-free. Proof: - -[XXX: will be updated once we are done massaging the lock_rename()] - First of all, at any moment we have a linear ordering of the - objects - A < B iff (A is an ancestor of B) or (B is not an ancestor - of A and ptr(A) < ptr(B)). - - That ordering can change. However, the following is true: - -(1) if object removal or non-cross-directory rename holds lock on A and - attempts to acquire lock on B, A will remain the parent of B until we - acquire the lock on B. (Proof: only cross-directory rename can change - the parent of object and it would have to lock the parent). - -(2) if cross-directory rename holds the lock on filesystem, order will not - change until rename acquires all locks. (Proof: other cross-directory - renames will be blocked on filesystem lock and we don't start changing - the order until we had acquired all locks). - -(3) locks on non-directory objects are acquired only after locks on - directory objects, and are acquired in inode pointer order. - (Proof: all operations but renames take lock on at most one - non-directory object, except renames, which take locks on source and - target in inode pointer order in the case they are not directories.) - -Now consider the minimal deadlock. Each process is blocked on -attempt to acquire some lock and already holds at least one lock. Let's -consider the set of contended locks. First of all, filesystem lock is -not contended, since any process blocked on it is not holding any locks. -Thus all processes are blocked on ->i_rwsem. - -By (3), any process holding a non-directory lock can only be -waiting on another non-directory lock with a larger address. Therefore -the process holding the "largest" such lock can always make progress, and -non-directory objects are not included in the set of contended locks. - -Thus link creation can't be a part of deadlock - it can't be -blocked on source and it means that it doesn't hold any locks. - -Any contended object is either held by cross-directory rename or -has a child that is also contended. Indeed, suppose that it is held by -operation other than cross-directory rename. Then the lock this operation -is blocked on belongs to child of that object due to (1). - -It means that one of the operations is cross-directory rename. -Otherwise the set of contended objects would be infinite - each of them -would have a contended child and we had assumed that no object is its -own descendent. Moreover, there is exactly one cross-directory rename -(see above). - -Consider the object blocking the cross-directory rename. One -of its descendents is locked by cross-directory rename (otherwise we -would again have an infinite set of contended objects). But that -means that cross-directory rename is taking locks out of order. Due -to (2) the order hadn't changed since we had acquired filesystem lock. -But locking rules for cross-directory rename guarantee that we do not -try to acquire lock on descendent before the lock on ancestor. -Contradiction. I.e. deadlock is impossible. Q.E.D. - - -These operations are guaranteed to avoid loop creation. Indeed, + There is a ranking on the locks, such that all primitives take +them in order of non-decreasing rank. Namely, + * rank ->i_rwsem of non-directories on given filesystem in inode +pointer order. + * put ->i_rwsem of all directories on a filesystem at the same rank, +lower than ->i_rwsem of any non-directory on the same filesystem. + * put ->s_vfs_rename_mutex at rank lower than that of any ->i_rwsem +on the same filesystem. + * among the locks on different filesystems use the relative +rank of those filesystems. + +For example, if we have NFS filesystem caching on a local one, we have + ->s_vfs_rename_mutex of NFS filesystem + ->i_rwsem of directories on that NFS filesystem, same rank for all + ->i_rwsem of non-directories on that filesystem, in order of +increasing address of inode + ->s_vfs_rename_mutex of local filesystem + ->i_rwsem of directories on the local filesystem, same rank for all + ->i_rwsem of non-directories on local filesystem, in order of +increasing address of inode. + + It's easy to verify that operations never take a lock with rank +lower than that of an already held lock. + + Suppose deadlocks are possible. Consider the minimal deadlocked +set of threads. It is a cycle of several threads, each blocked on a lock +held by the next thread in the cycle. + + Since the locking order is consistent with the ranking, all +contended locks in the minimal deadlock will be of the same rank, +i.e. they all will be ->i_rwsem of directories on the same filesystem. +Moreover, without loss of generality we can assume that all operations +are done directly to that filesystem and none of them has actually +reached the method call. + + In other words, we have a cycle of threads, T1,..., Tn, +and the same number of directories (D1,...,Dn) such that + T1 is blocked on D1 which is held by T2 + T2 is blocked on D2 which is held by T3 + ... + Tn is blocked on Dn which is held by T1. + + Each operation in the minimal cycle must have locked at least +one directory and blocked on attempt to lock another. That leaves +only 3 possible operations: directory removal (locks parent, then +child), same-directory rename killing a subdirectory (ditto) and +cross-directory rename of some sort. + + There must be a cross-directory rename in the set; indeed, +if all operations had been of the "lock parent, then child" sort +we would have Dn a parent of D1, which is a parent of D2, which is +a parent of D3, ..., which is a parent of Dn. Relationships couldn't +have changed since the moment directory locks had been acquired, +so they would all hold simultaneously at the deadlock time and +we would have a loop. + + Since all operations are on the same filesystem, there can't be +more than one cross-directory rename among them. Without loss of +generality we can assume that T1 is the one doing a cross-directory +rename and everything else is of the "lock parent, then child" sort. + + In other words, we have a cross-directory rename that locked +Dn and blocked on attempt to lock D1, which is a parent of D2, which is +a parent of D3, ..., which is a parent of Dn. Relationships between +D1,...,Dn all hold simultaneously at the deadlock time. Moreover, +cross-directory rename does not get to locking any directories until it +has acquired filesystem lock and verified that directories involved have +a common ancestor, which guarantees that ancestry relationships between +all of them had been stable. + + Consider the order in which directories are locked by the +cross-directory rename; parents first, then possibly their children. +Dn and D1 would have to be among those, with Dn locked before D1. +Which pair could it be? + It can't be the parents - indeed, since D1 is an ancestor of Dn, +it would be the first parent to be locked. Therefore at least one of the +children must be involved and thus neither of them could be a descendent +of another - otherwise the operation would not have progressed past +locking the parents. + It can't be a parent and its child; otherwise we would've had +a loop, since the parents are locked before the children, so the parent +would have to be a descendent of its child. + It can't be a parent and a child of another parent either. +Otherwise the child of the parent in question would've been a descendent +of another child. + That leaves only one possibility - namely, both Dn and D1 are +among the children, in some order. But that is also impossible, since +neither of the children is a descendent of another. + + That concludes the proof, since the set of operations with the +properties requiered for a minimal deadlock can not exist. + + Note that the check for having a common ancestor in cross-directory +rename is crucial - without it a deadlock would be possible. Indeed, +suppose the parents are initially in different trees; we would lock the +parent of source, then try to lock the parent of target, only to have +an unrelated lookup splice a distant ancestor of source to some distant +descendent of the parent of target. At that point we have cross-directory +rename holding the lock on parent of source and trying to lock its +distant ancestor. Add a bunch of rmdir() attempts on all directories +in between (all of those would fail with -ENOTEMPTY, had they ever gotten +the locks) and voila - we have a deadlock. + + Loop avoidance. + + These operations are guaranteed to avoid loop creation. Indeed, the only operation that could introduce loops is cross-directory rename. Since the only new (parent, child) pair added by rename() is (new parent, source), such loop would have to contain these objects and the rest of it @@ -133,8 +242,7 @@ would have to exist before rename(). I.e. at the moment of loop creation rename() responsible for that would be holding filesystem lock and new parent would have to be equal to or a descendent of source. But that means that new parent had been equal to or a descendent of source since the moment when -we had acquired filesystem lock and rename() would fail with -ELOOP in that -case. +we had acquired filesystem lock and rename() would fail in that case. While this locking scheme works for arbitrary DAGs, it relies on ability to check that directory is a descendent of another object. Current diff --git a/Documentation/filesystems/porting.rst b/Documentation/filesystems/porting.rst index 9100969e7de6..33cd56e2ca1a 100644 --- a/Documentation/filesystems/porting.rst +++ b/Documentation/filesystems/porting.rst @@ -1079,3 +1079,12 @@ On same-directory ->rename() the (tautological) update of .. is not protected by any locks; just don't do it if the old parent is the same as the new one. We really can't lock two subdirectories in same-directory rename - not without deadlocks. + +--- + +**mandatory** + +lock_rename() and lock_rename_child() may fail in cross-directory case, if +their arguments do not have a common ancestor. In that case ERR_PTR(-EXDEV) +is returned, with no locks taken. In-tree users updated; out-of-tree ones +would need to do so. diff --git a/fs/cachefiles/namei.c b/fs/cachefiles/namei.c index 7bf7a5fcc045..7ade836beb58 100644 --- a/fs/cachefiles/namei.c +++ b/fs/cachefiles/namei.c @@ -305,6 +305,8 @@ int cachefiles_bury_object(struct cachefiles_cache *cache, /* do the multiway lock magic */ trap = lock_rename(cache->graveyard, dir); + if (IS_ERR(trap)) + return PTR_ERR(trap); /* do some checks before getting the grave dentry */ if (rep->d_parent != dir || IS_DEADDIR(d_inode(rep))) { diff --git a/fs/ecryptfs/inode.c b/fs/ecryptfs/inode.c index a25dd3d20008..8efd20dc902b 100644 --- a/fs/ecryptfs/inode.c +++ b/fs/ecryptfs/inode.c @@ -599,6 +599,8 @@ ecryptfs_rename(struct mnt_idmap *idmap, struct inode *old_dir, target_inode = d_inode(new_dentry); trap = lock_rename(lower_old_dir_dentry, lower_new_dir_dentry); + if (IS_ERR(trap)) + return PTR_ERR(trap); dget(lower_new_dentry); rc = -EINVAL; if (lower_old_dentry->d_parent != lower_old_dir_dentry) diff --git a/fs/namei.c b/fs/namei.c index 29bafbdb44ca..6b0302ac80d1 100644 --- a/fs/namei.c +++ b/fs/namei.c @@ -3014,21 +3014,37 @@ static inline int may_create(struct mnt_idmap *idmap, return inode_permission(idmap, dir, MAY_WRITE | MAY_EXEC); } +// p1 != p2, both are on the same filesystem, ->s_vfs_rename_mutex is held static struct dentry *lock_two_directories(struct dentry *p1, struct dentry *p2) { - struct dentry *p; + struct dentry *p = p1, *q = p2, *r; - p = d_ancestor(p2, p1); - if (p) { + while ((r = p->d_parent) != p2 && r != p) + p = r; + if (r == p2) { + // p is a child of p2 and an ancestor of p1 or p1 itself inode_lock_nested(p2->d_inode, I_MUTEX_PARENT); inode_lock_nested(p1->d_inode, I_MUTEX_PARENT2); return p; } - - p = d_ancestor(p1, p2); - inode_lock_nested(p1->d_inode, I_MUTEX_PARENT); - inode_lock_nested(p2->d_inode, I_MUTEX_PARENT2); - return p; + // p is the root of connected component that contains p1 + // p2 does not occur on the path from p to p1 + while ((r = q->d_parent) != p1 && r != p && r != q) + q = r; + if (r == p1) { + // q is a child of p1 and an ancestor of p2 or p2 itself + inode_lock_nested(p1->d_inode, I_MUTEX_PARENT); + inode_lock_nested(p2->d_inode, I_MUTEX_PARENT2); + return q; + } else if (likely(r == p)) { + // both p2 and p1 are descendents of p + inode_lock_nested(p1->d_inode, I_MUTEX_PARENT); + inode_lock_nested(p2->d_inode, I_MUTEX_PARENT2); + return NULL; + } else { // no common ancestor at the time we'd been called + mutex_unlock(&p1->d_sb->s_vfs_rename_mutex); + return ERR_PTR(-EXDEV); + } } /* @@ -4947,6 +4963,10 @@ int do_renameat2(int olddfd, struct filename *from, int newdfd, retry_deleg: trap = lock_rename(new_path.dentry, old_path.dentry); + if (IS_ERR(trap)) { + error = PTR_ERR(trap); + goto exit_lock_rename; + } old_dentry = lookup_one_qstr_excl(&old_last, old_path.dentry, lookup_flags); @@ -5014,6 +5034,7 @@ int do_renameat2(int olddfd, struct filename *from, int newdfd, dput(old_dentry); exit3: unlock_rename(new_path.dentry, old_path.dentry); +exit_lock_rename: if (delegated_inode) { error = break_deleg_wait(&delegated_inode); if (!error) diff --git a/fs/nfsd/vfs.c b/fs/nfsd/vfs.c index fbbea7498f02..a99260c3f9bc 100644 --- a/fs/nfsd/vfs.c +++ b/fs/nfsd/vfs.c @@ -1813,6 +1813,10 @@ nfsd_rename(struct svc_rqst *rqstp, struct svc_fh *ffhp, char *fname, int flen, } trap = lock_rename(tdentry, fdentry); + if (IS_ERR(trap)) { + err = (rqstp->rq_vers == 2) ? nfserr_acces : nfserr_xdev; + goto out; + } err = fh_fill_pre_attrs(ffhp); if (err != nfs_ok) goto out_unlock; diff --git a/fs/overlayfs/copy_up.c b/fs/overlayfs/copy_up.c index 4382881b0709..e44dc5f66161 100644 --- a/fs/overlayfs/copy_up.c +++ b/fs/overlayfs/copy_up.c @@ -722,7 +722,7 @@ static int ovl_copy_up_workdir(struct ovl_copy_up_ctx *c) struct inode *inode; struct inode *udir = d_inode(c->destdir), *wdir = d_inode(c->workdir); struct path path = { .mnt = ovl_upper_mnt(ofs) }; - struct dentry *temp, *upper; + struct dentry *temp, *upper, *trap; struct ovl_cu_creds cc; int err; struct ovl_cattr cattr = { @@ -760,9 +760,11 @@ static int ovl_copy_up_workdir(struct ovl_copy_up_ctx *c) * If temp was moved, abort without the cleanup. */ ovl_start_write(c->dentry); - if (lock_rename(c->workdir, c->destdir) != NULL || - temp->d_parent != c->workdir) { + trap = lock_rename(c->workdir, c->destdir); + if (trap || temp->d_parent != c->workdir) { err = -EIO; + if (IS_ERR(trap)) + goto out; goto unlock; } else if (err) { goto cleanup; @@ -803,6 +805,7 @@ static int ovl_copy_up_workdir(struct ovl_copy_up_ctx *c) ovl_set_flag(OVL_WHITEOUTS, inode); unlock: unlock_rename(c->workdir, c->destdir); +out: ovl_end_write(c->dentry); return err; diff --git a/fs/overlayfs/dir.c b/fs/overlayfs/dir.c index aab3f5d93556..0f8b4a719237 100644 --- a/fs/overlayfs/dir.c +++ b/fs/overlayfs/dir.c @@ -1180,6 +1180,10 @@ static int ovl_rename(struct mnt_idmap *idmap, struct inode *olddir, } trap = lock_rename(new_upperdir, old_upperdir); + if (IS_ERR(trap)) { + err = PTR_ERR(trap); + goto out_revert_creds; + } olddentry = ovl_lookup_upper(ofs, old->d_name.name, old_upperdir, old->d_name.len); diff --git a/fs/overlayfs/super.c b/fs/overlayfs/super.c index a0967bb25003..fc3a6ff648bd 100644 --- a/fs/overlayfs/super.c +++ b/fs/overlayfs/super.c @@ -439,8 +439,10 @@ static bool ovl_workdir_ok(struct dentry *workdir, struct dentry *upperdir) bool ok = false; if (workdir != upperdir) { - ok = (lock_rename(workdir, upperdir) == NULL); - unlock_rename(workdir, upperdir); + struct dentry *trap = lock_rename(workdir, upperdir); + if (!IS_ERR(trap)) + unlock_rename(workdir, upperdir); + ok = (trap == NULL); } return ok; } diff --git a/fs/overlayfs/util.c b/fs/overlayfs/util.c index 50a201e9cd39..7b667345e673 100644 --- a/fs/overlayfs/util.c +++ b/fs/overlayfs/util.c @@ -1198,12 +1198,17 @@ void ovl_nlink_end(struct dentry *dentry) int ovl_lock_rename_workdir(struct dentry *workdir, struct dentry *upperdir) { + struct dentry *trap; + /* Workdir should not be the same as upperdir */ if (workdir == upperdir) goto err; /* Workdir should not be subdir of upperdir and vice versa */ - if (lock_rename(workdir, upperdir) != NULL) + trap = lock_rename(workdir, upperdir); + if (IS_ERR(trap)) + goto err; + if (trap) goto err_unlock; return 0; diff --git a/fs/smb/server/vfs.c b/fs/smb/server/vfs.c index c53dea5598fc..4cf8523ad038 100644 --- a/fs/smb/server/vfs.c +++ b/fs/smb/server/vfs.c @@ -708,6 +708,10 @@ int ksmbd_vfs_rename(struct ksmbd_work *work, const struct path *old_path, goto out2; trap = lock_rename_child(old_child, new_path.dentry); + if (IS_ERR(trap)) { + err = PTR_ERR(trap); + goto out_drop_write; + } old_parent = dget(old_child->d_parent); if (d_unhashed(old_child)) { @@ -770,6 +774,7 @@ int ksmbd_vfs_rename(struct ksmbd_work *work, const struct path *old_path, out3: dput(old_parent); unlock_rename(old_parent, new_path.dentry); +out_drop_write: mnt_drop_write(old_path->mnt); out2: path_put(&new_path);