[PATCH] workqueue: Do not warn when cancelling WQ_MEM_RECLAIM work from !WQ_MEM_RECLAIM worker

[Tvrtko Ursulin <tursulin@xxxxxxxxxx>]

From: Tvrtko Ursulin <tvrtko.ursulin@xxxxxxxxxx>

After commit
746ae46c1113 ("drm/sched: Mark scheduler work queues with WQ_MEM_RECLAIM")
amdgpu started seeing the following warning:

 [ ] workqueue: WQ_MEM_RECLAIM sdma0:drm_sched_run_job_work [gpu_sched] is flushing !WQ_MEM_RECLAIM events:amdgpu_device_delay_enable_gfx_off [amdgpu]
...
 [ ] Workqueue: sdma0 drm_sched_run_job_work [gpu_sched]
...
 [ ] Call Trace:
 [ ]  <TASK>
...
 [ ]  ? check_flush_dependency+0xf5/0x110
...
 [ ]  cancel_delayed_work_sync+0x6e/0x80
 [ ]  amdgpu_gfx_off_ctrl+0xab/0x140 [amdgpu]
 [ ]  amdgpu_ring_alloc+0x40/0x50 [amdgpu]
 [ ]  amdgpu_ib_schedule+0xf4/0x810 [amdgpu]
 [ ]  ? drm_sched_run_job_work+0x22c/0x430 [gpu_sched]
 [ ]  amdgpu_job_run+0xaa/0x1f0 [amdgpu]
 [ ]  drm_sched_run_job_work+0x257/0x430 [gpu_sched]
 [ ]  process_one_work+0x217/0x720
...
 [ ]  </TASK>

The intent of the verifcation done in check_flush_depedency is to ensure
forward progress during memory reclaim, by flagging cases when either a
memory reclaim process, or a memory reclaim work item is flushed from a
context not marked as memory reclaim safe.

This is correct when flushing, but when called from the
cancel(_delayed)_work_sync() paths it is a false positive because work is
either already running, or will not be running at all. Therefore
cancelling it is safe and we can relax the warning criteria by letting the
helper know of the calling context.

Signed-off-by: Tvrtko Ursulin <tvrtko.ursulin@xxxxxxxxxx>
Fixes: fca839c00a12 ("workqueue: warn if memory reclaim tries to flush !WQ_MEM_RECLAIM workqueue")
References: 746ae46c1113 ("drm/sched: Mark scheduler work queues with WQ_MEM_RECLAIM")
Cc: Tejun Heo <tj@xxxxxxxxxx>
Cc: Peter Zijlstra <peterz@xxxxxxxxxxxxx>
Cc: Lai Jiangshan <jiangshanlai@xxxxxxxxx>
Cc: Alex Deucher <alexander.deucher@xxxxxxx>
Cc: Christian König <christian.koenig@xxxxxxx
Cc: Matthew Brost <matthew.brost@xxxxxxxxx>
Cc: <stable@xxxxxxxxxxxxxxx> # v4.5+
---
 kernel/workqueue.c | 22 +++++++++++++---------
 1 file changed, 13 insertions(+), 9 deletions(-)

diff --git a/kernel/workqueue.c b/kernel/workqueue.c
index 9949ffad8df0..7abba81296cd 100644
--- a/kernel/workqueue.c
+++ b/kernel/workqueue.c
@@ -3680,23 +3680,27 @@ void workqueue_softirq_dead(unsigned int cpu)
  * check_flush_dependency - check for flush dependency sanity
  * @target_wq: workqueue being flushed
  * @target_work: work item being flushed (NULL for workqueue flushes)
+ * @from_cancel: are we called from the work cancel path
  *
  * %current is trying to flush the whole @target_wq or @target_work on it.
- * If @target_wq doesn't have %WQ_MEM_RECLAIM, verify that %current is not
- * reclaiming memory or running on a workqueue which doesn't have
- * %WQ_MEM_RECLAIM as that can break forward-progress guarantee leading to
- * a deadlock.
+ * If this is not the cancel path (which implies work being flushed is either
+ * already running, or will not be at all), check if @target_wq doesn't have
+ * %WQ_MEM_RECLAIM and verify that %current is not reclaiming memory or running
+ * on a workqueue which doesn't have %WQ_MEM_RECLAIM as that can break forward-
+ * progress guarantee leading to a deadlock.
  */
 static void check_flush_dependency(struct workqueue_struct *target_wq,
-				   struct work_struct *target_work)
+				   struct work_struct *target_work,
+				   bool from_cancel)
 {
-	work_func_t target_func = target_work ? target_work->func : NULL;
+	work_func_t target_func;
 	struct worker *worker;
 
-	if (target_wq->flags & WQ_MEM_RECLAIM)
+	if (from_cancel || target_wq->flags & WQ_MEM_RECLAIM)
 		return;
 
 	worker = current_wq_worker();
+	target_func = target_work ? target_work->func : NULL;
 
 	WARN_ONCE(current->flags & PF_MEMALLOC,
 		  "workqueue: PF_MEMALLOC task %d(%s) is flushing !WQ_MEM_RECLAIM %s:%ps",
@@ -3966,7 +3970,7 @@ void __flush_workqueue(struct workqueue_struct *wq)
 		list_add_tail(&this_flusher.list, &wq->flusher_overflow);
 	}
 
-	check_flush_dependency(wq, NULL);
+	check_flush_dependency(wq, NULL, false);
 
 	mutex_unlock(&wq->mutex);
 
@@ -4141,7 +4145,7 @@ static bool start_flush_work(struct work_struct *work, struct wq_barrier *barr,
 	}
 
 	wq = pwq->wq;
-	check_flush_dependency(wq, work);
+	check_flush_dependency(wq, work, from_cancel);
 
 	insert_wq_barrier(pwq, barr, work, worker);
 	raw_spin_unlock_irq(&pool->lock);
-- 
2.47.1