[PATCH 3.2 003/107] jbd2: protect all log tail updates with j_checkpoint_mutex

[Ben Hutchings]

3.2.72-rc1 review patch.  If anyone has any objections, please let me know.

------------------

From: Jan Kara <jack@xxxxxxx>

commit a78bb11d7acd525623c6a0c2ff4e213d527573fa upstream.

There are some log tail updates that are not protected by j_checkpoint_mutex.
Some of these are harmless because they happen during startup or shutdown but
updates in jbd2_journal_commit_transaction() and jbd2_journal_flush() can
really race with other log tail updates (e.g. someone doing
jbd2_journal_flush() with someone running jbd2_cleanup_journal_tail()). So
protect all log tail updates with j_checkpoint_mutex.

Signed-off-by: Jan Kara <jack@xxxxxxx>
Signed-off-by: "Theodore Ts'o" <tytso@xxxxxxx>
[bwh: Backported to 3.2:
 - Adjust context
 - Add unlock on the error path in jbd2_journal_flush()]
Signed-off-by: Ben Hutchings <ben@xxxxxxxxxxxxxxx>
Cc: Bartosz Kwitniewski <zerg2000@xxxxxxxxxxxxx>
---
 fs/jbd2/commit.c  |  2 ++
 fs/jbd2/journal.c | 19 ++++++++++++++++---
 2 files changed, 18 insertions(+), 3 deletions(-)

--- a/fs/jbd2/commit.c
+++ b/fs/jbd2/commit.c
@@ -340,6 +340,7 @@ void jbd2_journal_commit_transaction(jou
 	/* Do we need to erase the effects of a prior jbd2_journal_flush? */
 	if (journal->j_flags & JBD2_FLUSHED) {
 		jbd_debug(3, "super block updated\n");
+		mutex_lock(&journal->j_checkpoint_mutex);
 		/*
 		 * We hold j_checkpoint_mutex so tail cannot change under us.
 		 * We don't need any special data guarantees for writing sb
@@ -350,6 +351,7 @@ void jbd2_journal_commit_transaction(jou
 						journal->j_tail_sequence,
 						journal->j_tail,
 						WRITE_SYNC);
+		mutex_unlock(&journal->j_checkpoint_mutex);
 	} else {
 		jbd_debug(3, "superblock not updated\n");
 	}
--- a/fs/jbd2/journal.c
+++ b/fs/jbd2/journal.c
@@ -1242,6 +1242,8 @@ static int journal_reset(journal_t *jour
 			journal->j_errno);
 		journal->j_flags |= JBD2_FLUSHED;
 	} else {
+		/* Lock here to make assertions happy... */
+		mutex_lock(&journal->j_checkpoint_mutex);
 		/*
 		 * Update log tail information. We use WRITE_FUA since new
 		 * transaction will start reusing journal space and so we
@@ -1252,6 +1254,7 @@ static int journal_reset(journal_t *jour
 						journal->j_tail_sequence,
 						journal->j_tail,
 						WRITE_FUA);
+		mutex_unlock(&journal->j_checkpoint_mutex);
 	}
 	return jbd2_journal_start_thread(journal);
 }
@@ -1314,6 +1317,7 @@ int jbd2_journal_update_sb_log_tail(jour
 	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",
 		  tail_block, tail_tid);
 
@@ -1344,6 +1348,7 @@ static void jbd2_mark_journal_empty(jour
 {
 	journal_superblock_t *sb = journal->j_superblock;
 
+	BUG_ON(!mutex_is_locked(&journal->j_checkpoint_mutex));
 	read_lock(&journal->j_state_lock);
 	jbd_debug(1, "JBD2: Marking journal as empty (seq %d)\n",
 		  journal->j_tail_sequence);
@@ -1577,9 +1582,11 @@ int jbd2_journal_destroy(journal_t *jour
 	spin_unlock(&journal->j_list_lock);
 
 	if (journal->j_sb_buffer) {
-		if (!is_journal_aborted(journal))
+		if (!is_journal_aborted(journal)) {
+			mutex_lock(&journal->j_checkpoint_mutex);
 			jbd2_mark_journal_empty(journal);
-		else
+			mutex_unlock(&journal->j_checkpoint_mutex);
+		} else
 			err = -EIO;
 		brelse(journal->j_sb_buffer);
 	}
@@ -1828,10 +1835,13 @@ int jbd2_journal_flush(journal_t *journa
 	if (is_journal_aborted(journal))
 		return -EIO;
 
+	mutex_lock(&journal->j_checkpoint_mutex);
 	if (!err) {
 		err = jbd2_cleanup_journal_tail(journal);
-		if (err < 0)
+		if (err < 0) {
+			mutex_unlock(&journal->j_checkpoint_mutex);
 			goto out;
+		}
 		err = 0;
 	}
 
@@ -1841,6 +1851,7 @@ int jbd2_journal_flush(journal_t *journa
 	 * commits of data to the journal will restore the current
 	 * s_start value. */
 	jbd2_mark_journal_empty(journal);
+	mutex_unlock(&journal->j_checkpoint_mutex);
 	write_lock(&journal->j_state_lock);
 	J_ASSERT(!journal->j_running_transaction);
 	J_ASSERT(!journal->j_committing_transaction);
@@ -1882,8 +1893,12 @@ int jbd2_journal_wipe(journal_t *journal
 		write ? "Clearing" : "Ignoring");
 
 	err = jbd2_journal_skip_recovery(journal);
-	if (write)
+	if (write) {
+		/* Lock to make assertions happy... */
+		mutex_lock(&journal->j_checkpoint_mutex);
 		jbd2_mark_journal_empty(journal);
+		mutex_unlock(&journal->j_checkpoint_mutex);
+	}
 
  no_recovery:
 	return err;

--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/