From patchwork Sun Nov 13 21:21:08 2016 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: Richard Weinberger X-Patchwork-Id: 9424985 Return-Path: Received: from mail.wl.linuxfoundation.org (pdx-wl-mail.web.codeaurora.org [172.30.200.125]) by pdx-korg-patchwork.web.codeaurora.org (Postfix) with ESMTP id 62F266047D for ; Sun, 13 Nov 2016 21:25:34 +0000 (UTC) Received: from mail.wl.linuxfoundation.org (localhost [127.0.0.1]) by mail.wl.linuxfoundation.org (Postfix) with ESMTP id 53536288A2 for ; Sun, 13 Nov 2016 21:25:34 +0000 (UTC) Received: by mail.wl.linuxfoundation.org (Postfix, from userid 486) id 4600E28906; Sun, 13 Nov 2016 21:25:34 +0000 (UTC) X-Spam-Checker-Version: SpamAssassin 3.3.1 (2010-03-16) on pdx-wl-mail.web.codeaurora.org X-Spam-Level: X-Spam-Status: No, score=-6.9 required=2.0 tests=BAYES_00,RCVD_IN_DNSWL_HI autolearn=unavailable version=3.3.1 Received: from vger.kernel.org (vger.kernel.org [209.132.180.67]) by mail.wl.linuxfoundation.org (Postfix) with ESMTP id AB2EB288A2 for ; Sun, 13 Nov 2016 21:25:33 +0000 (UTC) Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S935289AbcKMVYT (ORCPT ); Sun, 13 Nov 2016 16:24:19 -0500 Received: from mail.sigma-star.at ([95.130.255.111]:46006 "EHLO mail.sigma-star.at" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S964934AbcKMVWJ (ORCPT ); Sun, 13 Nov 2016 16:22:09 -0500 Received: from localhost (localhost.localdomain [127.0.0.1]) by mail.sigma-star.at (Postfix) with ESMTP id 610CC24E001A; Sun, 13 Nov 2016 22:22:06 +0100 (CET) Received: from linux.site (richard.vpn.sigmapriv.at [10.3.0.5]) by mail.sigma-star.at (Postfix) with ESMTPSA id 242F324E0022; Sun, 13 Nov 2016 22:22:04 +0100 (CET) From: Richard Weinberger To: linux-mtd@lists.infradead.org Cc: linux-kernel@vger.kernel.org, linux-fsdevel@vger.kernel.org, dedekind1@gmail.com, adrian.hunter@intel.com, tytso@mit.edu, jaegeuk@kernel.org, david@sigma-star.at, wd@denx.de, sbabic@denx.de, dengler@linutronix.de, ebiggers@google.com, mhalcrow@google.com, hch@infradead.org, Richard Weinberger Subject: [PATCH 25/29] ubifs: Add full hash lookup support Date: Sun, 13 Nov 2016 22:21:08 +0100 Message-Id: <1479072072-6844-26-git-send-email-richard@nod.at> X-Mailer: git-send-email 2.7.3 In-Reply-To: <1479072072-6844-1-git-send-email-richard@nod.at> References: <1479072072-6844-1-git-send-email-richard@nod.at> Sender: linux-fsdevel-owner@vger.kernel.org Precedence: bulk List-ID: X-Mailing-List: linux-fsdevel@vger.kernel.org X-Virus-Scanned: ClamAV using ClamSMTP UBIFS stores a 32bit hash of every file, for traditional lookups by name this scheme is fine since UBIFS can first try to find the file by the hash of the filename and upon collisions it can walk through all entries with the same hash and do a string compare. When filesnames are encrypted fscrypto will ask the filesystem for a unique cookie, based on this cookie the filesystem has to be able to locate the target file again. With 32bit hashes this is impossible because the chance for collisions is very high. Do deal with that we store a 32bit cookie directly in the UBIFS directory entry node such that we get a 64bit cookie (32bit from filename hash and the dent cookie). For a lookup by hash UBIFS finds the entry by the first 32bit and then compares the dent cookie. If it does not match, it has to do a linear search of the whole directory and compares all dent cookies until the correct entry is found. Signed-off-by: Richard Weinberger --- fs/ubifs/dir.c | 7 +++- fs/ubifs/journal.c | 1 - fs/ubifs/tnc.c | 99 +++++++++++++++++++++++++++++++++++++++++++++++++- fs/ubifs/ubifs-media.h | 5 ++- fs/ubifs/ubifs.h | 2 + 5 files changed, 107 insertions(+), 7 deletions(-) diff --git a/fs/ubifs/dir.c b/fs/ubifs/dir.c index 7d3bc3fb8831..7f01f3d2ac3b 100644 --- a/fs/ubifs/dir.c +++ b/fs/ubifs/dir.c @@ -253,7 +253,7 @@ static struct dentry *ubifs_lookup(struct inode *dir, struct dentry *dentry, ubifs_assert(fname_len(&nm) == 0); ubifs_assert(fname_name(&nm) == NULL); dent_key_init_hash(c, &key, dir->i_ino, nm.hash); - err = ubifs_tnc_lookup(c, &key, dent); + err = ubifs_tnc_lookup_dh(c, &key, dent, nm.minor_hash); } else { dent_key_init(c, &key, dir->i_ino, &nm); err = ubifs_tnc_lookup_nm(c, &key, dent, &nm); @@ -628,7 +628,10 @@ static int ubifs_readdir(struct file *file, struct dir_context *ctx) if (encrypted) { fstr.len = fstr_real_len; - err = fscrypt_fname_disk_to_usr(dir, key_hash_flash(c, &dent->key), 0, &nm.disk_name, &fstr); + err = fscrypt_fname_disk_to_usr(dir, key_hash_flash(c, + &dent->key), + le32_to_cpu(dent->cookie), + &nm.disk_name, &fstr); if (err) goto out; } else { diff --git a/fs/ubifs/journal.c b/fs/ubifs/journal.c index f2b989dbe25a..0698cccc6223 100644 --- a/fs/ubifs/journal.c +++ b/fs/ubifs/journal.c @@ -78,7 +78,6 @@ static inline void zero_ino_node_unused(struct ubifs_ino_node *ino) static inline void zero_dent_node_unused(struct ubifs_dent_node *dent) { dent->padding1 = 0; - memset(dent->padding2, 0, 4); } /** diff --git a/fs/ubifs/tnc.c b/fs/ubifs/tnc.c index 0d751297873e..1eaf994addb4 100644 --- a/fs/ubifs/tnc.c +++ b/fs/ubifs/tnc.c @@ -1783,7 +1783,7 @@ int ubifs_tnc_bulk_read(struct ubifs_info *c, struct bu_info *bu) * @node: the node is returned here * @nm: node name * - * This function look up and reads a node which contains name hash in the key. + * This function looks up and reads a node which contains name hash in the key. * Since the hash may have collisions, there may be many nodes with the same * key, so we have to sequentially look to all of them until the needed one is * found. This function returns zero in case of success, %-ENOENT if the node @@ -1831,7 +1831,7 @@ static int do_lookup_nm(struct ubifs_info *c, const union ubifs_key *key, * @node: the node is returned here * @nm: node name * - * This function look up and reads a node which contains name hash in the key. + * This function looks up and reads a node which contains name hash in the key. * Since the hash may have collisions, there may be many nodes with the same * key, so we have to sequentially look to all of them until the needed one is * found. This function returns zero in case of success, %-ENOENT if the node @@ -1859,9 +1859,104 @@ int ubifs_tnc_lookup_nm(struct ubifs_info *c, const union ubifs_key *key, * Unluckily, there are hash collisions and we have to iterate over * them look at each direntry with colliding name hash sequentially. */ + return do_lookup_nm(c, key, node, nm); } +static int do_lookup_dh(struct ubifs_info *c, const union ubifs_key *key, + struct ubifs_dent_node *dent, uint32_t cookie) +{ + int n, err, type = key_type(c, key); + struct ubifs_znode *znode; + struct ubifs_zbranch *zbr; + union ubifs_key *cur_key, start_key; + + ubifs_assert(is_hash_key(c, key)); + + lowest_dent_key(c, &start_key, key_inum(c, key)); + cur_key = &start_key; + + for (;;) { + mutex_lock(&c->tnc_mutex); + + err = ubifs_lookup_level0(c, cur_key, &znode, &n); + if (unlikely(err < 0)) + goto out_unlock; + + if (!err) { + err = tnc_next(c, &znode, &n); + if (err) + goto out_unlock; + } + + zbr = &znode->zbranch[n]; + cur_key = &zbr->key; + + if (key_inum(c, cur_key) != key_inum(c, key) || + key_type(c, cur_key) != type) { + err = -ENOENT; + goto out_unlock; + } + + err = tnc_read_hashed_node(c, zbr, dent); + if (err) + goto out_unlock; + + if (key_hash(c, key) == key_hash(c, cur_key) && + le32_to_cpu(dent->cookie) == cookie) { + /* We found the entry. :-) */ + err = 0; + goto out_unlock; + } + + mutex_unlock(&c->tnc_mutex); + } + + ubifs_assert(0); + +out_unlock: + mutex_unlock(&c->tnc_mutex); + return err; +} + +/** + * ubifs_tnc_lookup_dh - look up a "double hashed" node. + * @c: UBIFS file-system description object + * @key: node key to lookup + * @node: the node is returned here + * @cookie: node cookie for collision resolution + * + * This function looks up and reads a node which contains name hash in the key. + * Since the hash may have collisions, there may be many nodes with the same + * key, so we have to sequentially look to all of them until the needed one + * with the same cookie value is found. + * This function returns zero in case of success, %-ENOENT if the node + * was not found, and a negative error code in case of failure. + */ +int ubifs_tnc_lookup_dh(struct ubifs_info *c, const union ubifs_key *key, + void *node, uint32_t cookie) +{ + int err; + const struct ubifs_dent_node *dent = node; + + /* + * We assume that in most of the cases there are no name collisions and + * 'ubifs_tnc_lookup()' returns us the right direntry. + */ + err = ubifs_tnc_lookup(c, key, node); + if (err) + return err; + + if (le32_to_cpu(dent->cookie) == cookie) + return 0; + + /* + * Unluckily, there are hash collisions and we have to iterate over + * them look at each direntry with colliding name hash sequentially. + */ + return do_lookup_dh(c, key, node, cookie); +} + /** * correct_parent_keys - correct parent znodes' keys. * @c: UBIFS file-system description object diff --git a/fs/ubifs/ubifs-media.h b/fs/ubifs/ubifs-media.h index e46331dcca4c..249124d9a801 100644 --- a/fs/ubifs/ubifs-media.h +++ b/fs/ubifs/ubifs-media.h @@ -530,7 +530,8 @@ struct ubifs_ino_node { * @padding1: reserved for future, zeroes * @type: type of the target inode (%UBIFS_ITYPE_REG, %UBIFS_ITYPE_DIR, etc) * @nlen: name length - * @padding2: reserved for future, zeroes + * @cookie: A 32bits random number, used to construct a 64bits + * identifier. * @name: zero-terminated name * * Note, do not forget to amend 'zero_dent_node_unused()' function when @@ -543,7 +544,7 @@ struct ubifs_dent_node { __u8 padding1; __u8 type; __le16 nlen; - __u8 padding2[4]; /* Watch 'zero_dent_node_unused()' if changing! */ + __le32 cookie; __u8 name[]; } __packed; diff --git a/fs/ubifs/ubifs.h b/fs/ubifs/ubifs.h index 122c7d95eb8e..8d9946082665 100644 --- a/fs/ubifs/ubifs.h +++ b/fs/ubifs/ubifs.h @@ -1580,6 +1580,8 @@ int ubifs_lookup_level0(struct ubifs_info *c, const union ubifs_key *key, struct ubifs_znode **zn, int *n); int ubifs_tnc_lookup_nm(struct ubifs_info *c, const union ubifs_key *key, void *node, const struct fscrypt_name *nm); +int ubifs_tnc_lookup_dh(struct ubifs_info *c, const union ubifs_key *key, + void *node, uint32_t secondary_hash); int ubifs_tnc_locate(struct ubifs_info *c, const union ubifs_key *key, void *node, int *lnum, int *offs); int ubifs_tnc_add(struct ubifs_info *c, const union ubifs_key *key, int lnum,