jbd2: Add sanity check for attempts to start handle during umount

An attempt to modify the file system during the call to
jbd2_destroy_journal() can lead to a system lockup.  So add some
checking to make it much more obvious when this happens to and to
determine where the offending code is located.

Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>

1 files changed, 10 insertions, 0 deletions

diff --git a/fs/jbd2/checkpoint.c b/fs/jbd2/checkpoint.c
index 5247e7ffdcb4..524800dce207 100644
--- a/fs/jbd2/checkpoint.c
+++ b/fs/jbd2/checkpoint.c
@@ -299,6 +299,16 @@ static int __process_buffer(journal_t *journal, struct journal_head *jh,
                 transaction->t_chp_stats.cs_forced_to_close++;
                 spin_unlock(&journal->j_list_lock);
                 jbd_unlock_bh_state(bh);
+                if (unlikely(journal->j_flags & JBD2_UNMOUNT))
+                        /*
+                         * The journal thread is dead; so starting and
+                         * waiting for a commit to finish will cause
+                         * us to wait for a _very_ long time.
+                         */
+                        printk(KERN_ERR "JBD2: %s: "
+                               "Waiting for Godot: block %llu\n",
+                               journal->j_devname,
+                               (unsigned long long) bh->b_blocknr);
                 jbd2_log_start_commit(journal, tid);
                 jbd2_log_wait_commit(journal, tid);
                 ret = 1;