diff --git a/drivers/gpu/host1x/bus.c b/drivers/gpu/host1x/bus.c
index 38f90ad0..01538262 100644
--- a/drivers/gpu/host1x/bus.c
+++ b/drivers/gpu/host1x/bus.c
@@ -32,6 +32,7 @@ struct host1x_subdev {
 /**
  * host1x_subdev_add() - add a new subdevice with an associated device node
  * @device: host1x device to add the subdevice to
+ * @driver: host1x driver containing the subdevices
  * @np: device node
  */
 static int host1x_subdev_add(struct host1x_device *device,
diff --git a/drivers/gpu/host1x/cdma.c b/drivers/gpu/host1x/cdma.c
index 8b9e7b4e..286e2f97 100644
--- a/drivers/gpu/host1x/cdma.c
+++ b/drivers/gpu/host1x/cdma.c
@@ -453,7 +453,7 @@ syncpt_incr:
 				u32 *mapped = cdma->push_buffer.mapped;
 
 				/*
-				 * Overwrite opcodes with 0 word writes to
+				 * Overwrite opcodes with 0 word writes
 				 * to offset 0xbad. This does nothing but
 				 * has a easily detected signature in debug
 				 * traces.
diff --git a/drivers/gpu/host1x/fence.c b/drivers/gpu/host1x/fence.c
index b2484606..e96ad93f 100644
--- a/drivers/gpu/host1x/fence.c
+++ b/drivers/gpu/host1x/fence.c
@@ -114,7 +114,7 @@ void host1x_fence_signal(struct host1x_syncpt_fence *f)
 	 */
 	cancel_delayed_work_sync(&f->timeout_work);
 
-	host1x_intr_put_ref(f->sp->host, f->sp->id, f->waiter_ref);
+	host1x_intr_put_ref(f->sp->host, f->sp->id, f->waiter_ref, false);
 
 	dma_fence_signal(&f->base);
 	dma_fence_put(&f->base);
@@ -133,7 +133,7 @@ static void do_fence_timeout(struct work_struct *work)
 	 * Cancel pending timeout work - if it races, it will
 	 * not get 'f->signaling' and return.
 	 */
-	host1x_intr_put_ref(f->sp->host, f->sp->id, f->waiter_ref);
+	host1x_intr_put_ref(f->sp->host, f->sp->id, f->waiter_ref, true);
 
 	dma_fence_set_error(&f->base, -ETIMEDOUT);
 	dma_fence_signal(&f->base);
diff --git a/drivers/gpu/host1x/intr.c b/drivers/gpu/host1x/intr.c
index 19b59c5c..bcffc4d7 100644
--- a/drivers/gpu/host1x/intr.c
+++ b/drivers/gpu/host1x/intr.c
@@ -251,7 +251,8 @@ int host1x_intr_add_action(struct host1x *host, struct host1x_syncpt *syncpt,
 	return 0;
 }
 
-void host1x_intr_put_ref(struct host1x *host, unsigned int id, void *ref)
+void host1x_intr_put_ref(struct host1x *host, unsigned int id, void *ref,
+			 bool flush)
 {
 	struct host1x_waitlist *waiter = ref;
 	struct host1x_syncpt *syncpt;
@@ -268,6 +269,12 @@ void host1x_intr_put_ref(struct host1x *host, unsigned int id, void *ref)
 	}
 	spin_unlock(&syncpt->intr.lock);
 
+	if (flush) {
+		/* Wait until any concurrently executing handler has finished. */
+		while (atomic_read(&waiter->state) != WLS_HANDLED)
+			cpu_relax();
+	}
+
 	kref_put(&waiter->refcount, waiter_release);
 }
 
diff --git a/drivers/gpu/host1x/intr.h b/drivers/gpu/host1x/intr.h
index dedbd0f7..e4c34609 100644
--- a/drivers/gpu/host1x/intr.h
+++ b/drivers/gpu/host1x/intr.h
@@ -76,8 +76,10 @@ int host1x_intr_add_action(struct host1x *host, struct host1x_syncpt *syncpt,
  * Unreference an action submitted to host1x_intr_add_action().
  * You must call this if you passed non-NULL as ref.
  * @ref the ref returned from host1x_intr_add_action()
+ * @flush wait until any pending handlers have completed before returning.
  */
-void host1x_intr_put_ref(struct host1x *host, unsigned int id, void *ref);
+void host1x_intr_put_ref(struct host1x *host, unsigned int id, void *ref,
+			 bool flush);
 
 /* Initialize host1x sync point interrupt */
 int host1x_intr_init(struct host1x *host, unsigned int irq_sync);
diff --git a/drivers/gpu/host1x/job.c b/drivers/gpu/host1x/job.c
index d7bb5d18..ae484d86 100644
--- a/drivers/gpu/host1x/job.c
+++ b/drivers/gpu/host1x/job.c
@@ -84,7 +84,7 @@ static void job_free(struct kref *ref)
 
 	if (job->waiter)
 		host1x_intr_put_ref(job->syncpt->host, job->syncpt->id,
-				    job->waiter);
+				    job->waiter, false);
 
 	if (job->syncpt)
 		host1x_syncpt_put(job->syncpt);
diff --git a/drivers/gpu/host1x/syncpt.c b/drivers/gpu/host1x/syncpt.c
index d0be7bdb..100270ac 100644
--- a/drivers/gpu/host1x/syncpt.c
+++ b/drivers/gpu/host1x/syncpt.c
@@ -295,7 +295,7 @@ int host1x_syncpt_wait(struct host1x_syncpt *sp, u32 thresh, long timeout,
 		}
 	}
 
-	host1x_intr_put_ref(sp->host, sp->id, ref);
+	host1x_intr_put_ref(sp->host, sp->id, ref, true);
 
 done:
 	return err;