ext4_fc_track_inode() may enter the wait loop with EXT4_STATE_FC_COMMITTING set but then observe it cleared before it actually sleeps. Asserting i_data_sem not held unconditionally triggers false lockdep warnings from paths that hold i_data_sem (e.g. fallocate) but never reach schedule(). Move lockdep_assert_not_held(&ei->i_data_sem) onto the schedule() path so it only fires when we are really going to sleep. Signed-off-by: Li Chen --- fs/ext4/fast_commit.c | 6 +++--- 1 file changed, 3 insertions(+), 3 deletions(-) diff --git a/fs/ext4/fast_commit.c b/fs/ext4/fast_commit.c index f3914f46bcf6..f24b4fea7421 100644 --- a/fs/ext4/fast_commit.c +++ b/fs/ext4/fast_commit.c @@ -577,8 +577,6 @@ void ext4_fc_track_inode(handle_t *handle, struct inode *inode) * commit. We shouldn't be holding i_data_sem when we go to sleep since * the commit path needs to grab the lock while committing the inode. */ - lockdep_assert_not_held(&ei->i_data_sem); - while (ext4_test_inode_state(inode, EXT4_STATE_FC_COMMITTING)) { #if (BITS_PER_LONG < 64) DEFINE_WAIT_BIT(wait, &ei->i_state_flags, @@ -592,8 +590,10 @@ void ext4_fc_track_inode(handle_t *handle, struct inode *inode) EXT4_STATE_FC_COMMITTING); #endif prepare_to_wait(wq, &wait.wq_entry, TASK_UNINTERRUPTIBLE); - if (ext4_test_inode_state(inode, EXT4_STATE_FC_COMMITTING)) + if (ext4_test_inode_state(inode, EXT4_STATE_FC_COMMITTING)) { + lockdep_assert_not_held(&ei->i_data_sem); schedule(); + } finish_wait(wq, &wait.wq_entry); } -- 2.51.0