[media] v4l: omap4iss: Reset the ISS when the pipeline can't be stopped

When a failure to stop a module in the pipeline is detected, the only
way to recover is to reset the ISS. However, as other users can be using
a different pipeline with other modules, the ISS can't be reset
synchronously with the error detection.
Keep track of modules that have failed to stop, and reset the ISS
accordingly when the last user releases the last reference to the ISS.
Refuse to start streaming on a pipeline that contains a crashed module,
as the hardware wouldn't work anyway.
Modify the omap4iss_pipeline_set_stream() function to record the new ISS
pipeline state only when no error occurs, except when stopping the
pipeline in which case the pipeline is still marked as stopped.

Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Signed-off-by: Mauro Carvalho Chehab <m.chehab@samsung.com>

1 files changed, 24 insertions, 0 deletions

diff --git a/drivers/staging/media/omap4iss/iss.c b/drivers/staging/media/omap4iss/iss.c
index dffa31e..5ad604d 100644
--- a/drivers/staging/media/omap4iss/iss.c
+++ b/drivers/staging/media/omap4iss/iss.c
@@ -573,12 +573,22 @@ static int iss_pipeline_link_notify(struct media_link *link, u32 flags,
 static int iss_pipeline_enable(struct iss_pipeline *pipe,
 			       enum iss_pipeline_stream_state mode)
 {
+	struct iss_device *iss = pipe->output->iss;
 	struct media_entity *entity;
 	struct media_pad *pad;
 	struct v4l2_subdev *subdev;
 	unsigned long flags;
 	int ret;
 
+	/* If one of the entities in the pipeline has crashed it will not work
+	 * properly. Refuse to start streaming in that case. This check must be
+	 * performed before the loop below to avoid starting entities if the
+	 * pipeline won't start anyway (those entities would then likely fail to
+	 * stop, making the problem worse).
+	 */
+	if (pipe->entities & iss->crashed)
+		return -EIO;
+
 	spin_lock_irqsave(&pipe->lock, flags);
 	pipe->state &= ~(ISS_PIPELINE_IDLE_INPUT | ISS_PIPELINE_IDLE_OUTPUT);
 	spin_unlock_irqrestore(&pipe->lock, flags);
@@ -617,6 +627,7 @@ static int iss_pipeline_enable(struct iss_pipeline *pipe,
  */
 static int iss_pipeline_disable(struct iss_pipeline *pipe)
 {
+	struct iss_device *iss = pipe->output->iss;
 	struct media_entity *entity;
 	struct media_pad *pad;
 	struct v4l2_subdev *subdev;
@@ -641,6 +652,11 @@ static int iss_pipeline_disable(struct iss_pipeline *pipe)
 		if (ret < 0) {
 			dev_dbg(iss->dev, "%s: module stop timeout.\n",
 				subdev->name);
+			/* If the entity failed to stopped, assume it has
+			 * crashed. Mark it as such, the ISS will be reset when
+			 * applications will release it.
+			 */
+			iss->crashed |= 1U << subdev->entity.id;
 			failure = -ETIMEDOUT;
 		}
 	}
@@ -715,6 +731,7 @@ static int iss_reset(struct iss_device *iss)
 		usleep_range(10, 10);
 	}
 
+	iss->crashed = 0;
 	return 0;
 }
 
@@ -1058,6 +1075,13 @@ void omap4iss_put(struct iss_device *iss)
 	BUG_ON(iss->ref_count == 0);
 	if (--iss->ref_count == 0) {
 		iss_disable_interrupts(iss);
+		/* Reset the ISS if an entity has failed to stop. This is the
+		 * only way to recover from such conditions, although it would
+		 * be worth investigating whether resetting the ISP only can't
+		 * fix the problem in some cases.
+		 */
+		if (iss->crashed)
+			iss_reset(iss);
 		iss_disable_clocks(iss);
 	}
 	mutex_unlock(&iss->iss_mutex);