Merge remote-tracking branch 'ubifs/linux-next'

8 files changed, 110 insertions, 33 deletions

diff --git a/fs/ubifs/commit.c b/fs/ubifs/commit.c
index 02429d81ca33..b148fbc80f8d 100644
--- a/fs/ubifs/commit.c
+++ b/fs/ubifs/commit.c
@@ -48,6 +48,56 @@
 #include <linux/slab.h>
 #include "ubifs.h"
 
+/*
+ * nothing_to_commit - check if there is nothing to commit.
+ * @c: UBIFS file-system description object
+ *
+ * This is a helper function which checks if there is anything to commit. It is
+ * used as an optimization to avoid starting the commit if it is not really
+ * necessary. Indeed, the commit operation always assumes flash I/O (e.g.,
+ * writing the commit start node to the log), and it is better to avoid doing
+ * this unnecessarily. E.g., 'ubifs_sync_fs()' runs the commit, but if there is
+ * nothing to commit, it is more optimal to avoid any flash I/O.
+ *
+ * This function has to be called with @c->commit_sem locked for writing -
+ * this function does not take LPT/TNC locks because the @c->commit_sem
+ * guarantees that we have exclusive access to the TNC and LPT data structures.
+ *
+ * This function returns %1 if there is nothing to commit and %0 otherwise.
+ */
+static int nothing_to_commit(struct ubifs_info *c)
+{
+	/*
+	 * During mounting or remounting from R/O mode to R/W mode we may
+	 * commit for various recovery-related reasons.
+	 */
+	if (c->mounting || c->remounting_rw)
+		return 0;
+
+	/*
+	 * If the root TNC node is dirty, we definitely have something to
+	 * commit.
+	 */
+	if (c->zroot.znode && test_bit(DIRTY_ZNODE, &c->zroot.znode->flags))
+		return 0;
+
+	/*
+	 * Even though the TNC is clean, the LPT tree may have dirty nodes. For
+	 * example, this may happen if the budgeting subsystem invoked GC to
+	 * make some free space, and the GC found an LEB with only dirty and
+	 * free space. In this case GC would just change the lprops of this
+	 * LEB (by turning all space into free space) and unmap it.
+	 */
+	if (c->nroot && test_bit(DIRTY_CNODE, &c->nroot->flags))
+		return 0;
+
+	ubifs_assert(atomic_long_read(&c->dirty_zn_cnt) == 0);
+	ubifs_assert(c->dirty_pn_cnt == 0);
+	ubifs_assert(c->dirty_nn_cnt == 0);
+
+	return 1;
+}
+
 /**
  * do_commit - commit the journal.
  * @c: UBIFS file-system description object
@@ -70,6 +120,12 @@ static int do_commit(struct ubifs_info *c)
 		goto out_up;
 	}
 
+	if (nothing_to_commit(c)) {
+		up_write(&c->commit_sem);
+		err = 0;
+		goto out_cancel;
+	}
+
 	/* Sync all write buffers (necessary for recovery) */
 	for (i = 0; i < c->jhead_cnt; i++) {
 		err = ubifs_wbuf_sync(&c->jheads[i].wbuf);
@@ -162,12 +218,12 @@ static int do_commit(struct ubifs_info *c)
 	if (err)
 		goto out;
 
+out_cancel:
 	spin_lock(&c->cs_lock);
 	c->cmt_state = COMMIT_RESTING;
 	wake_up(&c->cmt_wq);
 	dbg_cmt("commit end");
 	spin_unlock(&c->cs_lock);
-
 	return 0;
 
 out_up:
diff --git a/fs/ubifs/debug.c b/fs/ubifs/debug.c
index 0bee4dbffc31..bcb1acb79263 100644
--- a/fs/ubifs/debug.c
+++ b/fs/ubifs/debug.c
@@ -2813,19 +2813,19 @@ int dbg_debugfs_init_fs(struct ubifs_info *c)
 	}
 
 	fname = "dump_lprops";
-	dent = debugfs_create_file(fname, S_IWUGO, d->dfs_dir, c, &dfs_fops);
+	dent = debugfs_create_file(fname, S_IWUSR, d->dfs_dir, c, &dfs_fops);
 	if (IS_ERR(dent))
 		goto out_remove;
 	d->dfs_dump_lprops = dent;
 
 	fname = "dump_budg";
-	dent = debugfs_create_file(fname, S_IWUGO, d->dfs_dir, c, &dfs_fops);
+	dent = debugfs_create_file(fname, S_IWUSR, d->dfs_dir, c, &dfs_fops);
 	if (IS_ERR(dent))
 		goto out_remove;
 	d->dfs_dump_budg = dent;
 
 	fname = "dump_tnc";
-	dent = debugfs_create_file(fname, S_IWUGO, d->dfs_dir, c, &dfs_fops);
+	dent = debugfs_create_file(fname, S_IWUSR, d->dfs_dir, c, &dfs_fops);
 	if (IS_ERR(dent))
 		goto out_remove;
 	d->dfs_dump_tnc = dent;
diff --git a/fs/ubifs/io.c b/fs/ubifs/io.c
index d82173182eeb..d1fe56203a1d 100644
--- a/fs/ubifs/io.c
+++ b/fs/ubifs/io.c
@@ -88,8 +88,12 @@ void ubifs_ro_mode(struct ubifs_info *c, int err)
  * This function may skip data nodes CRC checking if @c->no_chk_data_crc is
  * true, which is controlled by corresponding UBIFS mount option. However, if
  * @must_chk_crc is true, then @c->no_chk_data_crc is ignored and CRC is
- * checked. Similarly, if @c->always_chk_crc is true, @c->no_chk_data_crc is
- * ignored and CRC is checked.
+ * checked. Similarly, if @c->mounting or @c->remounting_rw is true (we are
+ * mounting or re-mounting to R/W mode), @c->no_chk_data_crc is ignored and CRC
+ * is checked. This is because during mounting or re-mounting from R/O mode to
+ * R/W mode we may read journal nodes (when replying the journal or doing the
+ * recovery) and the journal nodes may potentially be corrupted, so checking is
+ * required.
  *
  * This function returns zero in case of success and %-EUCLEAN in case of bad
  * CRC or magic.
@@ -131,8 +135,8 @@ int ubifs_check_node(const struct ubifs_info *c, const void *buf, int lnum,
 		   node_len > c->ranges[type].max_len)
 		goto out_len;
 
-	if (!must_chk_crc && type == UBIFS_DATA_NODE && !c->always_chk_crc &&
-	     c->no_chk_data_crc)
+	if (!must_chk_crc && type == UBIFS_DATA_NODE && !c->mounting &&
+	    !c->remounting_rw && c->no_chk_data_crc)
 		return 0;
 
 	crc = crc32(UBIFS_CRC32_INIT, buf + 8, node_len - 8);
diff --git a/fs/ubifs/recovery.c b/fs/ubifs/recovery.c
index 77e9b874b6c2..e2714f8f05ff 100644
--- a/fs/ubifs/recovery.c
+++ b/fs/ubifs/recovery.c
@@ -28,6 +28,23 @@
  * UBIFS always cleans away all remnants of an unclean un-mount, so that
  * errors do not accumulate. However UBIFS defers recovery if it is mounted
  * read-only, and the flash is not modified in that case.
+ *
+ * The general UBIFS approach to the recovery is that it recovers from
+ * corruptions which could be caused by power cuts, but it refuses to recover
+ * from corruption caused by other reasons. And UBIFS tries to distinguish
+ * between these 2 reasons of corruptions and silently recover in the former
+ * case and loudly complain in the latter case.
+ *
+ * UBIFS writes only to erased LEBs, so it writes only to the flash space
+ * containing only 0xFFs. UBIFS also always writes strictly from the beginning
+ * of the LEB to the end. And UBIFS assumes that the underlying flash media
+ * writes in @c->min_io_unit bytes at a time.
+ *
+ * Hence, if UBIFS finds a corrupted node at offset X, it expects only the min.
+ * I/O unit corresponding to offset X to contain corrupted data, all the
+ * following min. I/O units have to contain empty space (all 0xFFs). If this is
+ * not true, the corruption cannot be the result of a power cut, and UBIFS
+ * refuses to mount.
  */
 
 #include <linux/crc32.h>
@@ -671,10 +688,14 @@ struct ubifs_scan_leb *ubifs_recover_leb(struct ubifs_info *c, int lnum,
 		} else {
 			int corruption = first_non_ff(buf, len);
 
+			/*
+			 * See header comment for this file for more
+			 * explanations about the reasons we have this check.
+			 */
 			ubifs_err("corrupt empty space LEB %d:%d, corruption "
 				  "starts at %d", lnum, offs, corruption);
 			/* Make sure we dump interesting non-0xFF data */
-			offs = corruption;
+			offs += corruption;
 			buf += corruption;
 			goto corrupted;
 		}
diff --git a/fs/ubifs/scan.c b/fs/ubifs/scan.c
index 3e1ee57dbeaa..36216b46f772 100644
--- a/fs/ubifs/scan.c
+++ b/fs/ubifs/scan.c
@@ -328,7 +328,7 @@ struct ubifs_scan_leb *ubifs_scan(const struct ubifs_info *c, int lnum,
 		if (!quiet)
 			ubifs_err("empty space starts at non-aligned offset %d",
 				  offs);
-		goto corrupted;;
+		goto corrupted;
 	}
 
 	ubifs_end_scan(c, sleb, lnum, offs);
diff --git a/fs/ubifs/super.c b/fs/ubifs/super.c
index 6e11c2975dcf..d203c99faa1f 100644
--- a/fs/ubifs/super.c
+++ b/fs/ubifs/super.c
@@ -1202,11 +1202,7 @@ static int mount_ubifs(struct ubifs_info *c)
 	if (c->bulk_read == 1)
 		bu_init(c);
 
-	/*
-	 * We have to check all CRCs, even for data nodes, when we mount the FS
-	 * (specifically, when we are replaying).
-	 */
-	c->always_chk_crc = 1;
+	c->mounting = 1;
 
 	err = ubifs_read_superblock(c);
 	if (err)
@@ -1382,7 +1378,7 @@ static int mount_ubifs(struct ubifs_info *c)
 	if (err)
 		goto out_infos;
 
-	c->always_chk_crc = 0;
+	c->mounting = 0;
 
 	ubifs_msg("mounted UBI device %d, volume %d, name \"%s\"",
 		  c->vi.ubi_num, c->vi.vol_id, c->vi.name);
@@ -1543,7 +1539,6 @@ static int ubifs_remount_rw(struct ubifs_info *c)
 	mutex_lock(&c->umount_mutex);
 	dbg_save_space_info(c);
 	c->remounting_rw = 1;
-	c->always_chk_crc = 1;
 
 	err = check_free_space(c);
 	if (err)
@@ -1650,7 +1645,6 @@ static int ubifs_remount_rw(struct ubifs_info *c)
 	dbg_gen("re-mounted read-write");
 	c->ro_mount = 0;
 	c->remounting_rw = 0;
-	c->always_chk_crc = 0;
 	err = dbg_check_space_info(c);
 	mutex_unlock(&c->umount_mutex);
 	return err;
@@ -1667,7 +1661,6 @@ out:
 	c->ileb_buf = NULL;
 	ubifs_lpt_free(c, 1);
 	c->remounting_rw = 0;
-	c->always_chk_crc = 0;
 	mutex_unlock(&c->umount_mutex);
 	return err;
 }
diff --git a/fs/ubifs/tnc.c b/fs/ubifs/tnc.c
index ad9cf0133622..de485979ca39 100644
--- a/fs/ubifs/tnc.c
+++ b/fs/ubifs/tnc.c
@@ -447,8 +447,11 @@ static int tnc_read_node_nm(struct ubifs_info *c, struct ubifs_zbranch *zbr,
  *
  * Note, this function does not check CRC of data nodes if @c->no_chk_data_crc
  * is true (it is controlled by corresponding mount option). However, if
- * @c->always_chk_crc is true, @c->no_chk_data_crc is ignored and CRC is always
- * checked.
+ * @c->mounting or @c->remounting_rw is true (we are mounting or re-mounting to
+ * R/W mode), @c->no_chk_data_crc is ignored and CRC is checked. This is
+ * because during mounting or re-mounting from R/O mode to R/W mode we may read
+ * journal nodes (when replying the journal or doing the recovery) and the
+ * journal nodes may potentially be corrupted, so checking is required.
  */
 static int try_read_node(const struct ubifs_info *c, void *buf, int type,
 			 int len, int lnum, int offs)
@@ -476,7 +479,8 @@ static int try_read_node(const struct ubifs_info *c, void *buf, int type,
 	if (node_len != len)
 		return 0;
 
-	if (type == UBIFS_DATA_NODE && !c->always_chk_crc && c->no_chk_data_crc)
+	if (type == UBIFS_DATA_NODE && c->no_chk_data_crc && !c->mounting &&
+	    !c->remounting_rw)
 		return 1;
 
 	crc = crc32(UBIFS_CRC32_INIT, buf + 8, node_len - 8);
diff --git a/fs/ubifs/ubifs.h b/fs/ubifs/ubifs.h
index 381d6b207a52..d1823541f987 100644
--- a/fs/ubifs/ubifs.h
+++ b/fs/ubifs/ubifs.h
@@ -1166,22 +1166,21 @@ struct ubifs_debug_info;
  * @rp_uid: reserved pool user ID
  * @rp_gid: reserved pool group ID
  *
- * @empty: if the UBI device is empty
+ * @empty: %1 if the UBI device is empty
+ * @need_recovery: %1 if the file-system needs recovery
+ * @replaying: %1 during journal replay
+ * @mounting: %1 while mounting
+ * @remounting_rw: %1 while re-mounting from R/O mode to R/W mode
  * @replay_tree: temporary tree used during journal replay
  * @replay_list: temporary list used during journal replay
  * @replay_buds: list of buds to replay
  * @cs_sqnum: sequence number of first node in the log (commit start node)
  * @replay_sqnum: sequence number of node currently being replayed
- * @need_recovery: file-system needs recovery
- * @replaying: set to %1 during journal replay
  * @unclean_leb_list: LEBs to recover when re-mounting R/O mounted FS to R/W
  *                    mode
  * @rcvrd_mst_node: recovered master node to write when re-mounting R/O mounted
  *                  FS to R/W mode
  * @size_tree: inode size information for recovery
- * @remounting_rw: set while re-mounting from R/O mode to R/W mode
- * @always_chk_crc: always check CRCs (while mounting and remounting to R/W
- *                  mode)
  * @mount_opts: UBIFS-specific mount options
  *
  * @dbg: debugging-related information
@@ -1402,19 +1401,19 @@ struct ubifs_info {
 	gid_t rp_gid;
 
 	/* The below fields are used only during mounting and re-mounting */
-	int empty;
+	unsigned int empty:1;
+	unsigned int need_recovery:1;
+	unsigned int replaying:1;
+	unsigned int mounting:1;
+	unsigned int remounting_rw:1;
 	struct rb_root replay_tree;
 	struct list_head replay_list;
 	struct list_head replay_buds;
 	unsigned long long cs_sqnum;
 	unsigned long long replay_sqnum;
-	int need_recovery;
-	int replaying;
 	struct list_head unclean_leb_list;
 	struct ubifs_mst_node *rcvrd_mst_node;
 	struct rb_root size_tree;
-	int remounting_rw;
-	int always_chk_crc;
 	struct ubifs_mount_opts mount_opts;
 
 #ifdef CONFIG_UBIFS_FS_DEBUG