From: Du, Changbin <changbin.du@intel.com>
Date: Thu, 27 Oct 2016 03:10:31 +0000 (+0800)
Subject: drm/i915/gvt: fix nested sleeping issue
X-Git-Url: https://git.stricted.de/?a=commitdiff_plain;h=e45d7b7f47a4849a5d3d55a2cf5802a72924d37b;p=GitHub%2Fmoto-9609%2Fandroid_kernel_motorola_exynos9610.git

drm/i915/gvt: fix nested sleeping issue

We cannot use blocking method mutex_lock inside a wait loop.
Here we invoke pick_next_workload() which needs acquire a
mutex in our "condition" experssion. Then we go into a another
of the going-to-sleep sequence and changing the task state.
This is a dangerous. Let's rewrite the wait sequence to avoid
nested sleeping.

v2: fix do...while loop exit condition (zhenyu)
v3: rebase to gvt-staging branch

Signed-off-by: Du, Changbin <changbin.du@intel.com>
Signed-off-by: Zhenyu Wang <zhenyuw@linux.intel.com>
---

diff --git a/drivers/gpu/drm/i915/gvt/scheduler.c b/drivers/gpu/drm/i915/gvt/scheduler.c
index e96eaeebeb0a..f7e320b9b17a 100644
--- a/drivers/gpu/drm/i915/gvt/scheduler.c
+++ b/drivers/gpu/drm/i915/gvt/scheduler.c
@@ -402,19 +402,24 @@ static int workload_thread(void *priv)
 	struct intel_vgpu_workload *workload = NULL;
 	int ret;
 	bool need_force_wake = IS_SKYLAKE(gvt->dev_priv);
+	DEFINE_WAIT_FUNC(wait, woken_wake_function);
 
 	kfree(p);
 
 	gvt_dbg_core("workload thread for ring %d started\n", ring_id);
 
 	while (!kthread_should_stop()) {
-		ret = wait_event_interruptible(scheduler->waitq[ring_id],
-				kthread_should_stop() ||
-				(workload = pick_next_workload(gvt, ring_id)));
-
-		WARN_ON_ONCE(ret);
-
-		if (kthread_should_stop())
+		add_wait_queue(&scheduler->waitq[ring_id], &wait);
+		do {
+			workload = pick_next_workload(gvt, ring_id);
+			if (workload)
+				break;
+			wait_woken(&wait, TASK_INTERRUPTIBLE,
+				   MAX_SCHEDULE_TIMEOUT);
+		} while (!kthread_should_stop());
+		remove_wait_queue(&scheduler->waitq[ring_id], &wait);
+
+		if (!workload)
 			break;
 
 		mutex_lock(&scheduler_mutex);