nilfs2: fix potential hang in nilfs_detach_log_writer()

commit eb85dace897c5986bc2f36b3c783c6abb8a4292e upstream.

Syzbot has reported a potential hang in nilfs_detach_log_writer() called
during nilfs2 unmount.

Analysis revealed that this is because nilfs_segctor_sync(), which
synchronizes with the log writer thread, can be called after
nilfs_segctor_destroy() terminates that thread, as shown in the call trace
below:

nilfs_detach_log_writer
  nilfs_segctor_destroy
    nilfs_segctor_kill_thread  --> Shut down log writer thread
    flush_work
      nilfs_iput_work_func
        nilfs_dispose_list
          iput
            nilfs_evict_inode
              nilfs_transaction_commit
                nilfs_construct_segment (if inode needs sync)
                  nilfs_segctor_sync  --> Attempt to synchronize with
                                          log writer thread
                           *** DEADLOCK ***

Fix this issue by changing nilfs_segctor_sync() so that the log writer
thread returns normally without synchronizing after it terminates, and by
forcing tasks that are already waiting to complete once after the thread
terminates.

The skipped inode metadata flushout will then be processed together in the
subsequent cleanup work in nilfs_segctor_destroy().

Link: https://lkml.kernel.org/r/20240520132621.4054-4-konishi.ryusuke@gmail.com
Signed-off-by: Ryusuke Konishi <konishi.ryusuke@gmail.com>
Reported-by: syzbot+e3973c409251e136fdd0@syzkaller.appspotmail.com
Closes: https://syzkaller.appspot.com/bug?extid=e3973c409251e136fdd0
Tested-by: Ryusuke Konishi <konishi.ryusuke@gmail.com>
Cc: <stable@vger.kernel.org>
Cc: "Bai, Shuangpeng" <sjb7183@psu.edu>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

1 files changed, 18 insertions, 3 deletions

diff --git a/fs/nilfs2/segment.c b/fs/nilfs2/segment.c
index 987ddd2d75a1..379f3636ede3 100644
--- a/fs/nilfs2/segment.c
+++ b/fs/nilfs2/segment.c
@@ -2229,6 +2229,14 @@ static int nilfs_segctor_sync(struct nilfs_sc_info *sci)
 	for (;;) {
 		set_current_state(TASK_INTERRUPTIBLE);
 
+		/*
+		 * Synchronize only while the log writer thread is alive.
+		 * Leave flushing out after the log writer thread exits to
+		 * the cleanup work in nilfs_segctor_destroy().
+		 */
+		if (!sci->sc_task)
+			break;
+
 		if (atomic_read(&wait_req.done)) {
 			err = wait_req.err;
 			break;
@@ -2244,7 +2252,7 @@ static int nilfs_segctor_sync(struct nilfs_sc_info *sci)
 	return err;
 }
 
-static void nilfs_segctor_wakeup(struct nilfs_sc_info *sci, int err)
+static void nilfs_segctor_wakeup(struct nilfs_sc_info *sci, int err, bool force)
 {
 	struct nilfs_segctor_wait_request *wrq, *n;
 	unsigned long flags;
@@ -2252,7 +2260,7 @@ static void nilfs_segctor_wakeup(struct nilfs_sc_info *sci, int err)
 	spin_lock_irqsave(&sci->sc_wait_request.lock, flags);
 	list_for_each_entry_safe(wrq, n, &sci->sc_wait_request.head, wq.entry) {
 		if (!atomic_read(&wrq->done) &&
-		    nilfs_cnt32_ge(sci->sc_seq_done, wrq->seq)) {
+		    (force || nilfs_cnt32_ge(sci->sc_seq_done, wrq->seq))) {
 			wrq->err = err;
 			atomic_set(&wrq->done, 1);
 		}
@@ -2392,7 +2400,7 @@ static void nilfs_segctor_notify(struct nilfs_sc_info *sci, int mode, int err)
 	if (mode == SC_LSEG_SR) {
 		sci->sc_state &= ~NILFS_SEGCTOR_COMMIT;
 		sci->sc_seq_done = sci->sc_seq_accepted;
-		nilfs_segctor_wakeup(sci, err);
+		nilfs_segctor_wakeup(sci, err, false);
 		sci->sc_flush_request = 0;
 	} else {
 		if (mode == SC_FLUSH_FILE)
@@ -2774,6 +2782,13 @@ static void nilfs_segctor_destroy(struct nilfs_sc_info *sci)
 		|| sci->sc_seq_request != sci->sc_seq_done);
 	spin_unlock(&sci->sc_state_lock);
 
+	/*
+	 * Forcibly wake up tasks waiting in nilfs_segctor_sync(), which can
+	 * be called from delayed iput() via nilfs_evict_inode() and can race
+	 * with the above log writer thread termination.
+	 */
+	nilfs_segctor_wakeup(sci, 0, true);
+
 	if (flush_work(&sci->sc_iput_work))
 		flag = true;