jbd2: fix ocfs2 corrupt when updating journal superblock fails

If updating journal superblock fails after journal data has been
flushed, the error is omitted and this will mislead the caller as a
normal case.  In ocfs2, the checkpoint will be treated successfully
and the other node can get the lock to update. Since the sb_start is
still pointing to the old log block, it will rewrite the journal data
during journal recovery by the other node. Thus the new updates will
be overwritten and ocfs2 corrupts.  So in above case we have to return
the error, and ocfs2_commit_cache will take care of the error and
prevent the other node to do update first.  And only after recovering
journal it can do the new updates.

The issue discussion mail can be found at:
https://oss.oracle.com/pipermail/ocfs2-devel/2015-June/010856.html
http://comments.gmane.org/gmane.comp.file-systems.ext4/48841

[ Fixed bug in patch which allowed a non-negative error return from
  jbd2_cleanup_journal_tail() to leak out of jbd2_fjournal_flush(); this
  was causing xfstests ext4/306 to fail. -- Ted ]

Reported-by: Yiwen Jiang <jiangyiwen@huawei.com>
Signed-off-by: Joseph Qi <joseph.qi@huawei.com>
Signed-off-by: Theodore Ts'o <tytso@mit.edu>
Tested-by: Yiwen Jiang <jiangyiwen@huawei.com>
Cc: Junxiao Bi <junxiao.bi@oracle.com>
Cc: stable@vger.kernel.org

2 files changed, 33 insertions, 10 deletions

diff --git a/fs/jbd2/checkpoint.c b/fs/jbd2/checkpoint.c
index 6b7b73a..4227dc4 100644
--- a/fs/jbd2/checkpoint.c
+++ b/fs/jbd2/checkpoint.c
@@ -390,7 +390,7 @@ int jbd2_cleanup_journal_tail(journal_t *journal)
 	unsigned long	blocknr;
 
 	if (is_journal_aborted(journal))
-		return 1;
+		return -EIO;
 
 	if (!jbd2_journal_get_log_tail(journal, &first_tid, &blocknr))
 		return 1;
@@ -407,8 +407,7 @@ int jbd2_cleanup_journal_tail(journal_t *journal)
 	if (journal->j_flags & JBD2_BARRIER)
 		blkdev_issue_flush(journal->j_fs_dev, GFP_NOFS, NULL);
 
-	__jbd2_update_log_tail(journal, first_tid, blocknr);
-	return 0;
+	return __jbd2_update_log_tail(journal, first_tid, blocknr);
 }
 
 
diff --git a/fs/jbd2/journal.c b/fs/jbd2/journal.c
index 303ccd9..5804466 100644
--- a/fs/jbd2/journal.c
+++ b/fs/jbd2/journal.c
@@ -876,9 +876,10 @@ int jbd2_journal_get_log_tail(journal_t *journal, tid_t *tid,
  *
  * Requires j_checkpoint_mutex
  */
-void __jbd2_update_log_tail(journal_t *journal, tid_t tid, unsigned long block)
+int __jbd2_update_log_tail(journal_t *journal, tid_t tid, unsigned long block)
 {
 	unsigned long freed;
+	int ret;
 
 	BUG_ON(!mutex_is_locked(&journal->j_checkpoint_mutex));
 
@@ -888,7 +889,10 @@ void __jbd2_update_log_tail(journal_t *journal, tid_t tid, unsigned long block)
 	 * space and if we lose sb update during power failure we'd replay
 	 * old transaction with possibly newly overwritten data.
 	 */
-	jbd2_journal_update_sb_log_tail(journal, tid, block, WRITE_FUA);
+	ret = jbd2_journal_update_sb_log_tail(journal, tid, block, WRITE_FUA);
+	if (ret)
+		goto out;
+
 	write_lock(&journal->j_state_lock);
 	freed = block - journal->j_tail;
 	if (block < journal->j_tail)
@@ -904,6 +908,9 @@ void __jbd2_update_log_tail(journal_t *journal, tid_t tid, unsigned long block)
 	journal->j_tail_sequence = tid;
 	journal->j_tail = block;
 	write_unlock(&journal->j_state_lock);
+
+out:
+	return ret;
 }
 
 /*
@@ -1322,7 +1329,7 @@ static int journal_reset(journal_t *journal)
 	return jbd2_journal_start_thread(journal);
 }
 
-static void jbd2_write_superblock(journal_t *journal, int write_op)
+static int jbd2_write_superblock(journal_t *journal, int write_op)
 {
 	struct buffer_head *bh = journal->j_sb_buffer;
 	journal_superblock_t *sb = journal->j_superblock;
@@ -1361,7 +1368,10 @@ static void jbd2_write_superblock(journal_t *journal, int write_op)
 		printk(KERN_ERR "JBD2: Error %d detected when updating "
 		       "journal superblock for %s.\n", ret,
 		       journal->j_devname);
+		jbd2_journal_abort(journal, ret);
 	}
+
+	return ret;
 }
 
 /**
@@ -1374,10 +1384,11 @@ static void jbd2_write_superblock(journal_t *journal, int write_op)
  * Update a journal's superblock information about log tail and write it to
  * disk, waiting for the IO to complete.
  */
-void jbd2_journal_update_sb_log_tail(journal_t *journal, tid_t tail_tid,
+int jbd2_journal_update_sb_log_tail(journal_t *journal, tid_t tail_tid,
 				     unsigned long tail_block, int write_op)
 {
 	journal_superblock_t *sb = journal->j_superblock;
+	int ret;
 
 	BUG_ON(!mutex_is_locked(&journal->j_checkpoint_mutex));
 	jbd_debug(1, "JBD2: updating superblock (start %lu, seq %u)\n",
@@ -1386,13 +1397,18 @@ void jbd2_journal_update_sb_log_tail(journal_t *journal, tid_t tail_tid,
 	sb->s_sequence = cpu_to_be32(tail_tid);
 	sb->s_start    = cpu_to_be32(tail_block);
 
-	jbd2_write_superblock(journal, write_op);
+	ret = jbd2_write_superblock(journal, write_op);
+	if (ret)
+		goto out;
 
 	/* Log is no longer empty */
 	write_lock(&journal->j_state_lock);
 	WARN_ON(!sb->s_sequence);
 	journal->j_flags &= ~JBD2_FLUSHED;
 	write_unlock(&journal->j_state_lock);
+
+out:
+	return ret;
 }
 
 /**
@@ -1941,7 +1957,14 @@ int jbd2_journal_flush(journal_t *journal)
 		return -EIO;
 
 	mutex_lock(&journal->j_checkpoint_mutex);
-	jbd2_cleanup_journal_tail(journal);
+	if (!err) {
+		err = jbd2_cleanup_journal_tail(journal);
+		if (err < 0) {
+			mutex_unlock(&journal->j_checkpoint_mutex);
+			goto out;
+		}
+		err = 0;
+	}
 
 	/* Finally, mark the journal as really needing no recovery.
 	 * This sets s_start==0 in the underlying superblock, which is
@@ -1957,7 +1980,8 @@ int jbd2_journal_flush(journal_t *journal)
 	J_ASSERT(journal->j_head == journal->j_tail);
 	J_ASSERT(journal->j_tail_sequence == journal->j_transaction_sequence);
 	write_unlock(&journal->j_state_lock);
-	return 0;
+out:
+	return err;
 }
 
 /**