Merge branch 'linux-next' of git://git.infradead.org/ubifs-2.6

* 'linux-next' of git://git.infradead.org/ubifs-2.6:
  UBIFS: fix-up free space earlier
  UBIFS: intialize LPT earlier
  UBIFS: assert no fixup when writing a node
  UBIFS: fix clean znode counter corruption in error cases
  UBIFS: fix memory leak on error path
  UBIFS: fix shrinker object count reports
  UBIFS: fix recovery broken by the previous recovery fix
  UBIFS: amend ubifs_recover_leb interface
  UBIFS: introduce a "grouped" journal head flag
  UBIFS: supress false error messages

9 files changed, 136 insertions, 97 deletions

diff --git a/fs/ubifs/io.c b/fs/ubifs/io.c
index 166951e..3be645e 100644
--- a/fs/ubifs/io.c
+++ b/fs/ubifs/io.c
@@ -581,6 +581,7 @@ int ubifs_wbuf_write_nolock(struct ubifs_wbuf *wbuf, void *buf, int len)
 	ubifs_assert(wbuf->size % c->min_io_size == 0);
 	ubifs_assert(mutex_is_locked(&wbuf->io_mutex));
 	ubifs_assert(!c->ro_media && !c->ro_mount);
+	ubifs_assert(!c->space_fixup);
 	if (c->leb_size - wbuf->offs >= c->max_write_size)
 		ubifs_assert(!((wbuf->offs + wbuf->size) % c->max_write_size));
 
@@ -759,6 +760,7 @@ int ubifs_write_node(struct ubifs_info *c, void *buf, int len, int lnum,
 	ubifs_assert(lnum >= 0 && lnum < c->leb_cnt && offs >= 0);
 	ubifs_assert(offs % c->min_io_size == 0 && offs < c->leb_size);
 	ubifs_assert(!c->ro_media && !c->ro_mount);
+	ubifs_assert(!c->space_fixup);
 
 	if (c->ro_error)
 		return -EROFS;
diff --git a/fs/ubifs/journal.c b/fs/ubifs/journal.c
index 34b1679..cef0460 100644
--- a/fs/ubifs/journal.c
+++ b/fs/ubifs/journal.c
@@ -669,6 +669,7 @@ out_free:
 
 out_release:
 	release_head(c, BASEHD);
+	kfree(dent);
 out_ro:
 	ubifs_ro_mode(c, err);
 	if (last_reference)
diff --git a/fs/ubifs/orphan.c b/fs/ubifs/orphan.c
index bd644bf..a5422ff 100644
--- a/fs/ubifs/orphan.c
+++ b/fs/ubifs/orphan.c
@@ -674,7 +674,7 @@ static int kill_orphans(struct ubifs_info *c)
 		if (IS_ERR(sleb)) {
 			if (PTR_ERR(sleb) == -EUCLEAN)
 				sleb = ubifs_recover_leb(c, lnum, 0,
-							 c->sbuf, 0);
+							 c->sbuf, -1);
 			if (IS_ERR(sleb)) {
 				err = PTR_ERR(sleb);
 				break;
diff --git a/fs/ubifs/recovery.c b/fs/ubifs/recovery.c
index 731d9e2..783d8e0 100644
--- a/fs/ubifs/recovery.c
+++ b/fs/ubifs/recovery.c
@@ -564,19 +564,15 @@ static int fix_unclean_leb(struct ubifs_info *c, struct ubifs_scan_leb *sleb,
 }
 
 /**
- * drop_last_node - drop the last node or group of nodes.
+ * drop_last_group - drop the last group of nodes.
  * @sleb: scanned LEB information
  * @offs: offset of dropped nodes is returned here
- * @grouped: non-zero if whole group of nodes have to be dropped
  *
  * This is a helper function for 'ubifs_recover_leb()' which drops the last
- * node of the scanned LEB or the last group of nodes if @grouped is not zero.
- * This function returns %1 if a node was dropped and %0 otherwise.
+ * group of nodes of the scanned LEB.
  */
-static int drop_last_node(struct ubifs_scan_leb *sleb, int *offs, int grouped)
+static void drop_last_group(struct ubifs_scan_leb *sleb, int *offs)
 {
-	int dropped = 0;
-
 	while (!list_empty(&sleb->nodes)) {
 		struct ubifs_scan_node *snod;
 		struct ubifs_ch *ch;
@@ -585,17 +581,40 @@ static int drop_last_node(struct ubifs_scan_leb *sleb, int *offs, int grouped)
 				  list);
 		ch = snod->node;
 		if (ch->group_type != UBIFS_IN_NODE_GROUP)
-			return dropped;
-		dbg_rcvry("dropping node at %d:%d", sleb->lnum, snod->offs);
+			break;
+
+		dbg_rcvry("dropping grouped node at %d:%d",
+			  sleb->lnum, snod->offs);
+		*offs = snod->offs;
+		list_del(&snod->list);
+		kfree(snod);
+		sleb->nodes_cnt -= 1;
+	}
+}
+
+/**
+ * drop_last_node - drop the last node.
+ * @sleb: scanned LEB information
+ * @offs: offset of dropped nodes is returned here
+ * @grouped: non-zero if whole group of nodes have to be dropped
+ *
+ * This is a helper function for 'ubifs_recover_leb()' which drops the last
+ * node of the scanned LEB.
+ */
+static void drop_last_node(struct ubifs_scan_leb *sleb, int *offs)
+{
+	struct ubifs_scan_node *snod;
+
+	if (!list_empty(&sleb->nodes)) {
+		snod = list_entry(sleb->nodes.prev, struct ubifs_scan_node,
+				  list);
+
+		dbg_rcvry("dropping last node at %d:%d", sleb->lnum, snod->offs);
 		*offs = snod->offs;
 		list_del(&snod->list);
 		kfree(snod);
 		sleb->nodes_cnt -= 1;
-		dropped = 1;
-		if (!grouped)
-			break;
 	}
-	return dropped;
 }
 
 /**
@@ -604,7 +623,8 @@ static int drop_last_node(struct ubifs_scan_leb *sleb, int *offs, int grouped)
  * @lnum: LEB number
  * @offs: offset
  * @sbuf: LEB-sized buffer to use
- * @grouped: nodes may be grouped for recovery
+ * @jhead: journal head number this LEB belongs to (%-1 if the LEB does not
+ *         belong to any journal head)
  *
  * This function does a scan of a LEB, but caters for errors that might have
  * been caused by the unclean unmount from which we are attempting to recover.
@@ -612,13 +632,14 @@ static int drop_last_node(struct ubifs_scan_leb *sleb, int *offs, int grouped)
  * found, and a negative error code in case of failure.
  */
 struct ubifs_scan_leb *ubifs_recover_leb(struct ubifs_info *c, int lnum,
-					 int offs, void *sbuf, int grouped)
+					 int offs, void *sbuf, int jhead)
 {
 	int ret = 0, err, len = c->leb_size - offs, start = offs, min_io_unit;
+	int grouped = jhead == -1 ? 0 : c->jheads[jhead].grouped;
 	struct ubifs_scan_leb *sleb;
 	void *buf = sbuf + offs;
 
-	dbg_rcvry("%d:%d", lnum, offs);
+	dbg_rcvry("%d:%d, jhead %d, grouped %d", lnum, offs, jhead, grouped);
 
 	sleb = ubifs_start_scan(c, lnum, offs, sbuf);
 	if (IS_ERR(sleb))
@@ -635,7 +656,7 @@ struct ubifs_scan_leb *ubifs_recover_leb(struct ubifs_info *c, int lnum,
 		 * Scan quietly until there is an error from which we cannot
 		 * recover
 		 */
-		ret = ubifs_scan_a_node(c, buf, len, lnum, offs, 0);
+		ret = ubifs_scan_a_node(c, buf, len, lnum, offs, 1);
 		if (ret == SCANNED_A_NODE) {
 			/* A valid node, and not a padding node */
 			struct ubifs_ch *ch = buf;
@@ -695,59 +716,62 @@ struct ubifs_scan_leb *ubifs_recover_leb(struct ubifs_info *c, int lnum,
 		 * If nodes are grouped, always drop the incomplete group at
 		 * the end.
 		 */
-		drop_last_node(sleb, &offs, 1);
+		drop_last_group(sleb, &offs);
 
-	/*
-	 * While we are in the middle of the same min. I/O unit keep dropping
-	 * nodes. So basically, what we want is to make sure that the last min.
-	 * I/O unit where we saw the corruption is dropped completely with all
-	 * the uncorrupted node which may possibly sit there.
-	 *
-	 * In other words, let's name the min. I/O unit where the corruption
-	 * starts B, and the previous min. I/O unit A. The below code tries to
-	 * deal with a situation when half of B contains valid nodes or the end
-	 * of a valid node, and the second half of B contains corrupted data or
-	 * garbage. This means that UBIFS had been writing to B just before the
-	 * power cut happened. I do not know how realistic is this scenario
-	 * that half of the min. I/O unit had been written successfully and the
-	 * other half not, but this is possible in our 'failure mode emulation'
-	 * infrastructure at least.
-	 *
-	 * So what is the problem, why we need to drop those nodes? Whey can't
-	 * we just clean-up the second half of B by putting a padding node
-	 * there? We can, and this works fine with one exception which was
-	 * reproduced with power cut emulation testing and happens extremely
-	 * rarely. The description follows, but it is worth noting that that is
-	 * only about the GC head, so we could do this trick only if the bud
-	 * belongs to the GC head, but it does not seem to be worth an
-	 * additional "if" statement.
-	 *
-	 * So, imagine the file-system is full, we run GC which is moving valid
-	 * nodes from LEB X to LEB Y (obviously, LEB Y is the current GC head
-	 * LEB). The @c->gc_lnum is -1, which means that GC will retain LEB X
-	 * and will try to continue. Imagine that LEB X is currently the
-	 * dirtiest LEB, and the amount of used space in LEB Y is exactly the
-	 * same as amount of free space in LEB X.
-	 *
-	 * And a power cut happens when nodes are moved from LEB X to LEB Y. We
-	 * are here trying to recover LEB Y which is the GC head LEB. We find
-	 * the min. I/O unit B as described above. Then we clean-up LEB Y by
-	 * padding min. I/O unit. And later 'ubifs_rcvry_gc_commit()' function
-	 * fails, because it cannot find a dirty LEB which could be GC'd into
-	 * LEB Y! Even LEB X does not match because the amount of valid nodes
-	 * there does not fit the free space in LEB Y any more! And this is
-	 * because of the padding node which we added to LEB Y. The
-	 * user-visible effect of this which I once observed and analysed is
-	 * that we cannot mount the file-system with -ENOSPC error.
-	 *
-	 * So obviously, to make sure that situation does not happen we should
-	 * free min. I/O unit B in LEB Y completely and the last used min. I/O
-	 * unit in LEB Y should be A. This is basically what the below code
-	 * tries to do.
-	 */
-	while (min_io_unit == round_down(offs, c->min_io_size) &&
-	       min_io_unit != offs &&
-	       drop_last_node(sleb, &offs, grouped));
+	if (jhead == GCHD) {
+		/*
+		 * If this LEB belongs to the GC head then while we are in the
+		 * middle of the same min. I/O unit keep dropping nodes. So
+		 * basically, what we want is to make sure that the last min.
+		 * I/O unit where we saw the corruption is dropped completely
+		 * with all the uncorrupted nodes which may possibly sit there.
+		 *
+		 * In other words, let's name the min. I/O unit where the
+		 * corruption starts B, and the previous min. I/O unit A. The
+		 * below code tries to deal with a situation when half of B
+		 * contains valid nodes or the end of a valid node, and the
+		 * second half of B contains corrupted data or garbage. This
+		 * means that UBIFS had been writing to B just before the power
+		 * cut happened. I do not know how realistic is this scenario
+		 * that half of the min. I/O unit had been written successfully
+		 * and the other half not, but this is possible in our 'failure
+		 * mode emulation' infrastructure at least.
+		 *
+		 * So what is the problem, why we need to drop those nodes? Why
+		 * can't we just clean-up the second half of B by putting a
+		 * padding node there? We can, and this works fine with one
+		 * exception which was reproduced with power cut emulation
+		 * testing and happens extremely rarely.
+		 *
+		 * Imagine the file-system is full, we run GC which starts
+		 * moving valid nodes from LEB X to LEB Y (obviously, LEB Y is
+		 * the current GC head LEB). The @c->gc_lnum is -1, which means
+		 * that GC will retain LEB X and will try to continue. Imagine
+		 * that LEB X is currently the dirtiest LEB, and the amount of
+		 * used space in LEB Y is exactly the same as amount of free
+		 * space in LEB X.
+		 *
+		 * And a power cut happens when nodes are moved from LEB X to
+		 * LEB Y. We are here trying to recover LEB Y which is the GC
+		 * head LEB. We find the min. I/O unit B as described above.
+		 * Then we clean-up LEB Y by padding min. I/O unit. And later
+		 * 'ubifs_rcvry_gc_commit()' function fails, because it cannot
+		 * find a dirty LEB which could be GC'd into LEB Y! Even LEB X
+		 * does not match because the amount of valid nodes there does
+		 * not fit the free space in LEB Y any more! And this is
+		 * because of the padding node which we added to LEB Y. The
+		 * user-visible effect of this which I once observed and
+		 * analysed is that we cannot mount the file-system with
+		 * -ENOSPC error.
+		 *
+		 * So obviously, to make sure that situation does not happen we
+		 * should free min. I/O unit B in LEB Y completely and the last
+		 * used min. I/O unit in LEB Y should be A. This is basically
+		 * what the below code tries to do.
+		 */
+		while (offs > min_io_unit)
+			drop_last_node(sleb, &offs);
+	}
 
 	buf = sbuf + offs;
 	len = c->leb_size - offs;
@@ -881,7 +905,7 @@ struct ubifs_scan_leb *ubifs_recover_log_leb(struct ubifs_info *c, int lnum,
 		}
 		ubifs_scan_destroy(sleb);
 	}
-	return ubifs_recover_leb(c, lnum, offs, sbuf, 0);
+	return ubifs_recover_leb(c, lnum, offs, sbuf, -1);
 }
 
 /**
diff --git a/fs/ubifs/replay.c b/fs/ubifs/replay.c
index 6617280..5e97161 100644
--- a/fs/ubifs/replay.c
+++ b/fs/ubifs/replay.c
@@ -557,8 +557,7 @@ static int replay_bud(struct ubifs_info *c, struct bud_entry *b)
 		 * these LEBs could possibly be written to at the power cut
 		 * time.
 		 */
-		sleb = ubifs_recover_leb(c, lnum, offs, c->sbuf,
-					 b->bud->jhead != GCHD);
+		sleb = ubifs_recover_leb(c, lnum, offs, c->sbuf, b->bud->jhead);
 	else
 		sleb = ubifs_scan(c, lnum, offs, c->sbuf, 0);
 	if (IS_ERR(sleb))
diff --git a/fs/ubifs/shrinker.c b/fs/ubifs/shrinker.c
index ca953a9..9e1d056 100644
--- a/fs/ubifs/shrinker.c
+++ b/fs/ubifs/shrinker.c
@@ -284,7 +284,11 @@ int ubifs_shrinker(struct shrinker *shrink, struct shrink_control *sc)
 	long clean_zn_cnt = atomic_long_read(&ubifs_clean_zn_cnt);
 
 	if (nr == 0)
-		return clean_zn_cnt;
+		/*
+		 * Due to the way UBIFS updates the clean znode counter it may
+		 * temporarily be negative.
+		 */
+		return clean_zn_cnt >= 0 ? clean_zn_cnt : 1;
 
 	if (!clean_zn_cnt) {
 		/*
diff --git a/fs/ubifs/super.c b/fs/ubifs/super.c
index 1ab0d22..b5aeb5a 100644
--- a/fs/ubifs/super.c
+++ b/fs/ubifs/super.c
@@ -811,15 +811,18 @@ static int alloc_wbufs(struct ubifs_info *c)
 
 		c->jheads[i].wbuf.sync_callback = &bud_wbuf_callback;
 		c->jheads[i].wbuf.jhead = i;
+		c->jheads[i].grouped = 1;
 	}
 
 	c->jheads[BASEHD].wbuf.dtype = UBI_SHORTTERM;
 	/*
 	 * Garbage Collector head likely contains long-term data and
-	 * does not need to be synchronized by timer.
+	 * does not need to be synchronized by timer. Also GC head nodes are
+	 * not grouped.
 	 */
 	c->jheads[GCHD].wbuf.dtype = UBI_LONGTERM;
 	c->jheads[GCHD].wbuf.no_timer = 1;
+	c->jheads[GCHD].grouped = 0;
 
 	return 0;
 }
@@ -1284,12 +1287,25 @@ static int mount_ubifs(struct ubifs_info *c)
 	if ((c->mst_node->flags & cpu_to_le32(UBIFS_MST_DIRTY)) != 0) {
 		ubifs_msg("recovery needed");
 		c->need_recovery = 1;
-		if (!c->ro_mount) {
-			err = ubifs_recover_inl_heads(c, c->sbuf);
-			if (err)
-				goto out_master;
-		}
-	} else if (!c->ro_mount) {
+	}
+
+	if (c->need_recovery && !c->ro_mount) {
+		err = ubifs_recover_inl_heads(c, c->sbuf);
+		if (err)
+			goto out_master;
+	}
+
+	err = ubifs_lpt_init(c, 1, !c->ro_mount);
+	if (err)
+		goto out_master;
+
+	if (!c->ro_mount && c->space_fixup) {
+		err = ubifs_fixup_free_space(c);
+		if (err)
+			goto out_master;
+	}
+
+	if (!c->ro_mount) {
 		/*
 		 * Set the "dirty" flag so that if we reboot uncleanly we
 		 * will notice this immediately on the next mount.
@@ -1297,13 +1313,9 @@ static int mount_ubifs(struct ubifs_info *c)
 		c->mst_node->flags |= cpu_to_le32(UBIFS_MST_DIRTY);
 		err = ubifs_write_master(c);
 		if (err)
-			goto out_master;
+			goto out_lpt;
 	}
 
-	err = ubifs_lpt_init(c, 1, !c->ro_mount);
-	if (err)
-		goto out_lpt;
-
 	err = dbg_check_idx_size(c, c->bi.old_idx_sz);
 	if (err)
 		goto out_lpt;
@@ -1396,12 +1408,6 @@ static int mount_ubifs(struct ubifs_info *c)
 	} else
 		ubifs_assert(c->lst.taken_empty_lebs > 0);
 
-	if (!c->ro_mount && c->space_fixup) {
-		err = ubifs_fixup_free_space(c);
-		if (err)
-			goto out_infos;
-	}
-
 	err = dbg_check_filesystem(c);
 	if (err)
 		goto out_infos;
diff --git a/fs/ubifs/tnc.c b/fs/ubifs/tnc.c
index 8119b1f..91b4213 100644
--- a/fs/ubifs/tnc.c
+++ b/fs/ubifs/tnc.c
@@ -2876,12 +2876,13 @@ static void tnc_destroy_cnext(struct ubifs_info *c)
  */
 void ubifs_tnc_close(struct ubifs_info *c)
 {
-	long clean_freed;
-
 	tnc_destroy_cnext(c);
 	if (c->zroot.znode) {
-		clean_freed = ubifs_destroy_tnc_subtree(c->zroot.znode);
-		atomic_long_sub(clean_freed, &ubifs_clean_zn_cnt);
+		long n;
+
+		ubifs_destroy_tnc_subtree(c->zroot.znode);
+		n = atomic_long_read(&c->clean_zn_cnt);
+		atomic_long_sub(n, &ubifs_clean_zn_cnt);
 	}
 	kfree(c->gap_lebs);
 	kfree(c->ilebs);
diff --git a/fs/ubifs/ubifs.h b/fs/ubifs/ubifs.h
index a70d7b4..f79983d 100644
--- a/fs/ubifs/ubifs.h
+++ b/fs/ubifs/ubifs.h
@@ -722,12 +722,14 @@ struct ubifs_bud {
  * struct ubifs_jhead - journal head.
  * @wbuf: head's write-buffer
  * @buds_list: list of bud LEBs belonging to this journal head
+ * @grouped: non-zero if UBIFS groups nodes when writing to this journal head
  *
  * Note, the @buds list is protected by the @c->buds_lock.
  */
 struct ubifs_jhead {
 	struct ubifs_wbuf wbuf;
 	struct list_head buds_list;
+	unsigned int grouped:1;
 };
 
 /**
@@ -1742,7 +1744,7 @@ struct inode *ubifs_iget(struct super_block *sb, unsigned long inum);
 int ubifs_recover_master_node(struct ubifs_info *c);
 int ubifs_write_rcvrd_mst_node(struct ubifs_info *c);
 struct ubifs_scan_leb *ubifs_recover_leb(struct ubifs_info *c, int lnum,
-					 int offs, void *sbuf, int grouped);
+					 int offs, void *sbuf, int jhead);
 struct ubifs_scan_leb *ubifs_recover_log_leb(struct ubifs_info *c, int lnum,
 					     int offs, void *sbuf);
 int ubifs_recover_inl_heads(const struct ubifs_info *c, void *sbuf);