From d6e89786bed977f37f55ffca11e563f6d2b1e3b5 Mon Sep 17 00:00:00 2001 From: Johannes Berg Date: Wed, 22 Aug 2018 11:49:03 +0200 Subject: workqueue: skip lockdep wq dependency in cancel_work_sync() In cancel_work_sync(), we can only have one of two cases, even with an ordered workqueue: * the work isn't running, just cancelled before it started * the work is running, but then nothing else can be on the workqueue before it Thus, we need to skip the lockdep workqueue dependency handling, otherwise we get false positive reports from lockdep saying that we have a potential deadlock when the workqueue also has other work items with locking, e.g. work1_function() { mutex_lock(&mutex); ... } work2_function() { /* nothing */ } other_function() { queue_work(ordered_wq, &work1); queue_work(ordered_wq, &work2); mutex_lock(&mutex); cancel_work_sync(&work2); } As described above, this isn't a problem, but lockdep will currently flag it as if cancel_work_sync() was flush_work(), which *is* a problem. Signed-off-by: Johannes Berg Signed-off-by: Tejun Heo --- kernel/workqueue.c | 37 ++++++++++++++++++++++--------------- 1 file changed, 22 insertions(+), 15 deletions(-) diff --git a/kernel/workqueue.c b/kernel/workqueue.c index 7ea75529eabb..aa520e715bbc 100644 --- a/kernel/workqueue.c +++ b/kernel/workqueue.c @@ -2843,7 +2843,8 @@ reflush: } EXPORT_SYMBOL_GPL(drain_workqueue); -static bool start_flush_work(struct work_struct *work, struct wq_barrier *barr) +static bool start_flush_work(struct work_struct *work, struct wq_barrier *barr, + bool from_cancel) { struct worker *worker = NULL; struct worker_pool *pool; @@ -2885,7 +2886,8 @@ static bool start_flush_work(struct work_struct *work, struct wq_barrier *barr) * workqueues the deadlock happens when the rescuer stalls, blocking * forward progress. */ - if (pwq->wq->saved_max_active == 1 || pwq->wq->rescuer) { + if (!from_cancel && + (pwq->wq->saved_max_active == 1 || pwq->wq->rescuer)) { lock_map_acquire(&pwq->wq->lockdep_map); lock_map_release(&pwq->wq->lockdep_map); } @@ -2896,6 +2898,22 @@ already_gone: return false; } +static bool __flush_work(struct work_struct *work, bool from_cancel) +{ + struct wq_barrier barr; + + if (WARN_ON(!wq_online)) + return false; + + if (start_flush_work(work, &barr, from_cancel)) { + wait_for_completion(&barr.done); + destroy_work_on_stack(&barr.work); + return true; + } else { + return false; + } +} + /** * flush_work - wait for a work to finish executing the last queueing instance * @work: the work to flush @@ -2909,18 +2927,7 @@ already_gone: */ bool flush_work(struct work_struct *work) { - struct wq_barrier barr; - - if (WARN_ON(!wq_online)) - return false; - - if (start_flush_work(work, &barr)) { - wait_for_completion(&barr.done); - destroy_work_on_stack(&barr.work); - return true; - } else { - return false; - } + return __flush_work(work, false); } EXPORT_SYMBOL_GPL(flush_work); @@ -2986,7 +2993,7 @@ static bool __cancel_work_timer(struct work_struct *work, bool is_dwork) * isn't executing. */ if (wq_online) - flush_work(work); + __flush_work(work, true); clear_work_data(work); -- cgit v1.2.3 From 87915adc3f0acdf03c776df42e308e5a155c19af Mon Sep 17 00:00:00 2001 From: Johannes Berg Date: Wed, 22 Aug 2018 11:49:04 +0200 Subject: workqueue: re-add lockdep dependencies for flushing In flush_work(), we need to create a lockdep dependency so that the following scenario is appropriately tagged as a problem: work_function() { mutex_lock(&mutex); ... } other_function() { mutex_lock(&mutex); flush_work(&work); // or cancel_work_sync(&work); } This is a problem since the work might be running and be blocked on trying to acquire the mutex. Similarly, in flush_workqueue(). These were removed after cross-release partially caught these problems, but now cross-release was reverted anyway. IMHO the removal was erroneous anyway though, since lockdep should be able to catch potential problems, not just actual ones, and cross-release would only have caught the problem when actually invoking wait_for_completion(). Fixes: fd1a5b04dfb8 ("workqueue: Remove now redundant lock acquisitions wrt. workqueue flushes") Signed-off-by: Johannes Berg Signed-off-by: Tejun Heo --- kernel/workqueue.c | 8 ++++++++ 1 file changed, 8 insertions(+) diff --git a/kernel/workqueue.c b/kernel/workqueue.c index aa520e715bbc..661184fcd503 100644 --- a/kernel/workqueue.c +++ b/kernel/workqueue.c @@ -2652,6 +2652,9 @@ void flush_workqueue(struct workqueue_struct *wq) if (WARN_ON(!wq_online)) return; + lock_map_acquire(&wq->lockdep_map); + lock_map_release(&wq->lockdep_map); + mutex_lock(&wq->mutex); /* @@ -2905,6 +2908,11 @@ static bool __flush_work(struct work_struct *work, bool from_cancel) if (WARN_ON(!wq_online)) return false; + if (!from_cancel) { + lock_map_acquire(&work->lockdep_map); + lock_map_release(&work->lockdep_map); + } + if (start_flush_work(work, &barr, from_cancel)) { wait_for_completion(&barr.done); destroy_work_on_stack(&barr.work); -- cgit v1.2.3