diff mbox series

[v2,07/12] drm/i915: Protect workers against disappearing connectors

Message ID 20191212190230.188505-8-sean@poorly.run (mailing list archive)
State New, archived
Headers show
Series drm/i915: Add support for HDCP 1.4 over MST connectors | expand

Commit Message

Sean Paul Dec. 12, 2019, 7:02 p.m. UTC
From: Sean Paul <seanpaul@chromium.org>

This patch adds some protection against connectors being destroyed
before the HDCP workers are finished.

For check_work, we do a synchronous cancel after the connector is
unregistered which will ensure that it is finished before destruction.

In the case of prop_work, we can't do a synchronous wait since it needs
to take connection_mutex which could cause deadlock. Instead, we'll take
a reference on the connector when scheduling prop_work and give it up
once we're done.

Signed-off-by: Sean Paul <seanpaul@chromium.org>

Changes in v2:
- Added to the set
---
 drivers/gpu/drm/i915/display/intel_hdcp.c | 38 ++++++++++++++++++++---
 1 file changed, 33 insertions(+), 5 deletions(-)

Comments

Ramalingam C Dec. 13, 2019, 11:10 a.m. UTC | #1
On 2019-12-12 at 14:02:25 -0500, Sean Paul wrote:
> From: Sean Paul <seanpaul@chromium.org>
> 
> This patch adds some protection against connectors being destroyed
> before the HDCP workers are finished.
> 
> For check_work, we do a synchronous cancel after the connector is
> unregistered which will ensure that it is finished before destruction.
> 
> In the case of prop_work, we can't do a synchronous wait since it needs
> to take connection_mutex which could cause deadlock. Instead, we'll take
> a reference on the connector when scheduling prop_work and give it up
> once we're done.
> 
> Signed-off-by: Sean Paul <seanpaul@chromium.org>
Will there be an instance where prop_work is scheduled but before
execution cancelled from the queue itself? This will leak the connector
reference.

Atleast hdcp stack is not requesting for such action. So Looks good to me.

Reviewed-by: Ramalingam C <ramalingam.c@intel.com>
> 
> Changes in v2:
> - Added to the set
> ---
>  drivers/gpu/drm/i915/display/intel_hdcp.c | 38 ++++++++++++++++++++---
>  1 file changed, 33 insertions(+), 5 deletions(-)
> 
> diff --git a/drivers/gpu/drm/i915/display/intel_hdcp.c b/drivers/gpu/drm/i915/display/intel_hdcp.c
> index 798e7e1a19fc..c79dca2c74d1 100644
> --- a/drivers/gpu/drm/i915/display/intel_hdcp.c
> +++ b/drivers/gpu/drm/i915/display/intel_hdcp.c
> @@ -863,8 +863,10 @@ static void intel_hdcp_update_value(struct intel_connector *connector,
>  		return;
>  
>  	hdcp->value = value;
> -	if (update_property)
> +	if (update_property) {
> +		drm_connector_get(&connector->base);
>  		schedule_work(&hdcp->prop_work);
> +	}
>  }
>  
>  /* Implements Part 3 of the HDCP authorization procedure */
> @@ -954,6 +956,8 @@ static void intel_hdcp_prop_work(struct work_struct *work)
>  
>  	mutex_unlock(&hdcp->mutex);
>  	drm_modeset_unlock(&dev->mode_config.connection_mutex);
> +
> +	drm_connector_put(&connector->base);
>  }
>  
>  bool is_hdcp_supported(struct drm_i915_private *dev_priv, enum port port)
> @@ -1802,6 +1806,9 @@ static void intel_hdcp_check_work(struct work_struct *work)
>  					       check_work);
>  	struct intel_connector *connector = intel_hdcp_to_connector(hdcp);
>  
> +	if (drm_connector_is_unregistered(&connector->base))
> +		return;
> +
>  	if (!intel_hdcp2_check_link(connector))
>  		schedule_delayed_work(&hdcp->check_work,
>  				      DRM_HDCP2_CHECK_PERIOD_MS);
> @@ -2076,12 +2083,33 @@ void intel_hdcp_component_fini(struct drm_i915_private *dev_priv)
>  
>  void intel_hdcp_cleanup(struct intel_connector *connector)
>  {
> -	if (!connector->hdcp.shim)
> +	struct intel_hdcp *hdcp = &connector->hdcp;
> +
> +	if (!hdcp->shim)
>  		return;
>  
> -	mutex_lock(&connector->hdcp.mutex);
> -	kfree(connector->hdcp.port_data.streams);
> -	mutex_unlock(&connector->hdcp.mutex);
> +	WARN_ON(!drm_connector_is_unregistered(&connector->base));
> +
> +	/*
> +	 * Now that the connector is unregistered, check_work won't be run, but
> +	 * cancel any outstanding instances of it
> +	 */
> +	cancel_delayed_work_sync(&hdcp->check_work);
> +
> +	/*
> +	 * We don't cancel prop_work in the same way as check_work since it
> +	 * requires connection_mutex which could be held while calling this
> +	 * function. Instead, we rely on the connector references grabbed before
> +	 * scheduling prop_work to ensure the connector is alive when prop_work
> +	 * is run. So if we're in the destroy path (which is where this
> +	 * function should be called), we're "guaranteed" that prop_work is not
> +	 * active (tl;dr This Should Never Happen).
> +	 */
> +	WARN_ON(work_pending(&hdcp->prop_work));
> +
> +	mutex_lock(&hdcp->mutex);
> +	kfree(hdcp->port_data.streams);
> +	mutex_unlock(&hdcp->mutex);
>  }
>  
>  void intel_hdcp_atomic_check(struct drm_connector *connector,
> -- 
> Sean Paul, Software Engineer, Google / Chromium OS
>
Sean Paul Dec. 13, 2019, 7:04 p.m. UTC | #2
On Fri, Dec 13, 2019 at 04:40:33PM +0530, Ramalingam C wrote:
> On 2019-12-12 at 14:02:25 -0500, Sean Paul wrote:
> > From: Sean Paul <seanpaul@chromium.org>
> > 
> > This patch adds some protection against connectors being destroyed
> > before the HDCP workers are finished.
> > 
> > For check_work, we do a synchronous cancel after the connector is
> > unregistered which will ensure that it is finished before destruction.
> > 
> > In the case of prop_work, we can't do a synchronous wait since it needs
> > to take connection_mutex which could cause deadlock. Instead, we'll take
> > a reference on the connector when scheduling prop_work and give it up
> > once we're done.
> > 
> > Signed-off-by: Sean Paul <seanpaul@chromium.org>
> Will there be an instance where prop_work is scheduled but before
> execution cancelled from the queue itself? This will leak the connector
> reference.

No, prop_work is really quite simple, it just grabs some locks and updates the
property value. 

> 
> Atleast hdcp stack is not requesting for such action. So Looks good to me.
> 
> Reviewed-by: Ramalingam C <ramalingam.c@intel.com>

Thanks, I'm going to dig into what we should do when hdcp_cleanup is called from
connector_init failure paths and revise this patch.

> > 
> > Changes in v2:
> > - Added to the set
> > ---
> >  drivers/gpu/drm/i915/display/intel_hdcp.c | 38 ++++++++++++++++++++---
> >  1 file changed, 33 insertions(+), 5 deletions(-)
> > 
> > diff --git a/drivers/gpu/drm/i915/display/intel_hdcp.c b/drivers/gpu/drm/i915/display/intel_hdcp.c
> > index 798e7e1a19fc..c79dca2c74d1 100644
> > --- a/drivers/gpu/drm/i915/display/intel_hdcp.c
> > +++ b/drivers/gpu/drm/i915/display/intel_hdcp.c
> > @@ -863,8 +863,10 @@ static void intel_hdcp_update_value(struct intel_connector *connector,
> >  		return;
> >  
> >  	hdcp->value = value;
> > -	if (update_property)
> > +	if (update_property) {
> > +		drm_connector_get(&connector->base);
> >  		schedule_work(&hdcp->prop_work);
> > +	}
> >  }
> >  
> >  /* Implements Part 3 of the HDCP authorization procedure */
> > @@ -954,6 +956,8 @@ static void intel_hdcp_prop_work(struct work_struct *work)
> >  
> >  	mutex_unlock(&hdcp->mutex);
> >  	drm_modeset_unlock(&dev->mode_config.connection_mutex);
> > +
> > +	drm_connector_put(&connector->base);
> >  }
> >  
> >  bool is_hdcp_supported(struct drm_i915_private *dev_priv, enum port port)
> > @@ -1802,6 +1806,9 @@ static void intel_hdcp_check_work(struct work_struct *work)
> >  					       check_work);
> >  	struct intel_connector *connector = intel_hdcp_to_connector(hdcp);
> >  
> > +	if (drm_connector_is_unregistered(&connector->base))
> > +		return;
> > +
> >  	if (!intel_hdcp2_check_link(connector))
> >  		schedule_delayed_work(&hdcp->check_work,
> >  				      DRM_HDCP2_CHECK_PERIOD_MS);
> > @@ -2076,12 +2083,33 @@ void intel_hdcp_component_fini(struct drm_i915_private *dev_priv)
> >  
> >  void intel_hdcp_cleanup(struct intel_connector *connector)
> >  {
> > -	if (!connector->hdcp.shim)
> > +	struct intel_hdcp *hdcp = &connector->hdcp;
> > +
> > +	if (!hdcp->shim)
> >  		return;
> >  
> > -	mutex_lock(&connector->hdcp.mutex);
> > -	kfree(connector->hdcp.port_data.streams);
> > -	mutex_unlock(&connector->hdcp.mutex);
> > +	WARN_ON(!drm_connector_is_unregistered(&connector->base));
> > +
> > +	/*
> > +	 * Now that the connector is unregistered, check_work won't be run, but
> > +	 * cancel any outstanding instances of it
> > +	 */
> > +	cancel_delayed_work_sync(&hdcp->check_work);
> > +
> > +	/*
> > +	 * We don't cancel prop_work in the same way as check_work since it
> > +	 * requires connection_mutex which could be held while calling this
> > +	 * function. Instead, we rely on the connector references grabbed before
> > +	 * scheduling prop_work to ensure the connector is alive when prop_work
> > +	 * is run. So if we're in the destroy path (which is where this
> > +	 * function should be called), we're "guaranteed" that prop_work is not
> > +	 * active (tl;dr This Should Never Happen).
> > +	 */
> > +	WARN_ON(work_pending(&hdcp->prop_work));
> > +
> > +	mutex_lock(&hdcp->mutex);
> > +	kfree(hdcp->port_data.streams);
> > +	mutex_unlock(&hdcp->mutex);
> >  }
> >  
> >  void intel_hdcp_atomic_check(struct drm_connector *connector,
> > -- 
> > Sean Paul, Software Engineer, Google / Chromium OS
> >
diff mbox series

Patch

diff --git a/drivers/gpu/drm/i915/display/intel_hdcp.c b/drivers/gpu/drm/i915/display/intel_hdcp.c
index 798e7e1a19fc..c79dca2c74d1 100644
--- a/drivers/gpu/drm/i915/display/intel_hdcp.c
+++ b/drivers/gpu/drm/i915/display/intel_hdcp.c
@@ -863,8 +863,10 @@  static void intel_hdcp_update_value(struct intel_connector *connector,
 		return;
 
 	hdcp->value = value;
-	if (update_property)
+	if (update_property) {
+		drm_connector_get(&connector->base);
 		schedule_work(&hdcp->prop_work);
+	}
 }
 
 /* Implements Part 3 of the HDCP authorization procedure */
@@ -954,6 +956,8 @@  static void intel_hdcp_prop_work(struct work_struct *work)
 
 	mutex_unlock(&hdcp->mutex);
 	drm_modeset_unlock(&dev->mode_config.connection_mutex);
+
+	drm_connector_put(&connector->base);
 }
 
 bool is_hdcp_supported(struct drm_i915_private *dev_priv, enum port port)
@@ -1802,6 +1806,9 @@  static void intel_hdcp_check_work(struct work_struct *work)
 					       check_work);
 	struct intel_connector *connector = intel_hdcp_to_connector(hdcp);
 
+	if (drm_connector_is_unregistered(&connector->base))
+		return;
+
 	if (!intel_hdcp2_check_link(connector))
 		schedule_delayed_work(&hdcp->check_work,
 				      DRM_HDCP2_CHECK_PERIOD_MS);
@@ -2076,12 +2083,33 @@  void intel_hdcp_component_fini(struct drm_i915_private *dev_priv)
 
 void intel_hdcp_cleanup(struct intel_connector *connector)
 {
-	if (!connector->hdcp.shim)
+	struct intel_hdcp *hdcp = &connector->hdcp;
+
+	if (!hdcp->shim)
 		return;
 
-	mutex_lock(&connector->hdcp.mutex);
-	kfree(connector->hdcp.port_data.streams);
-	mutex_unlock(&connector->hdcp.mutex);
+	WARN_ON(!drm_connector_is_unregistered(&connector->base));
+
+	/*
+	 * Now that the connector is unregistered, check_work won't be run, but
+	 * cancel any outstanding instances of it
+	 */
+	cancel_delayed_work_sync(&hdcp->check_work);
+
+	/*
+	 * We don't cancel prop_work in the same way as check_work since it
+	 * requires connection_mutex which could be held while calling this
+	 * function. Instead, we rely on the connector references grabbed before
+	 * scheduling prop_work to ensure the connector is alive when prop_work
+	 * is run. So if we're in the destroy path (which is where this
+	 * function should be called), we're "guaranteed" that prop_work is not
+	 * active (tl;dr This Should Never Happen).
+	 */
+	WARN_ON(work_pending(&hdcp->prop_work));
+
+	mutex_lock(&hdcp->mutex);
+	kfree(hdcp->port_data.streams);
+	mutex_unlock(&hdcp->mutex);
 }
 
 void intel_hdcp_atomic_check(struct drm_connector *connector,