[PATCH v2 0/6] Fix legacy DPMS changes with MST

[PATCH v2 0/6] Fix legacy DPMS changes with MST

[Lyude Paul]

There's two major things this patchset does:
 - Add drm_dp_mst_connector_atomic_check() so drivers don't need to use
   ->best_encoder() to prevent modesets on zombie MST connectors. We'll
   use this later for implementing MST fallback retraining as well.
 - Fix DPMS on->off changes failing with legacy modesetting users after
   an MST connector's topology has disappeared, which resulted in CRTCs
   being left on when they shouldn't have been

Lyude Paul (6):
  drm/dp_mst: Introduce drm_dp_mst_connector_atomic_check()
  drm/nouveau: Unbreak nv50_mstc->best_encoder()
  drm/i915: Leave intel_conn->mst_port set, use mst_port_gone instead
  drm/i915: Skip vcpi allocation for MSTB ports that are gone
  drm/i915: Fix intel_dp_mst_best_encoder()
  drm/amdgpu/dm/mst: Use drm_dp_mst_connector_atomic_check()

 .../display/amdgpu_dm/amdgpu_dm_mst_types.c   | 12 +++
 drivers/gpu/drm/drm_dp_mst_topology.c         | 76 +++++++++++++++++++
 drivers/gpu/drm/i915/intel_dp_mst.c           | 46 ++++++-----
 drivers/gpu/drm/i915/intel_drv.h              |  1 +
 drivers/gpu/drm/nouveau/dispnv50/disp.c       | 25 +++---
 include/drm/drm_dp_mst_helper.h               |  3 +
 6 files changed, 132 insertions(+), 31 deletions(-)

-- 
2.17.1

[PATCH v2 1/6] drm/dp_mst: Introduce drm_dp_mst_connector_atomic_check()

[Lyude Paul]

Currently the way that we prevent userspace from performing new modesets
on MST connectors that have just been destroyed is rather broken.
There's nothing in the actual DRM DP MST topology helpers that checks
whether or not a connector still exists, instead each DRM driver does
this on it's own, usually by returning NULL from the best_encoder
callback which in turn, causes the atomic commit to fail.

However, this is wrong in a rather subtle way. If ->best_encoder()
returns NULL, this makes ALL modesets involving the connector fail. This
includes modesets from userspace that would shut off the CRTCs being
used by the connector. Since this results in blocking any changes to a
connector's DPMS prop, it has the sideaffect of preventing legacy
modesetting users from ever disabling a CRTC that was previously enabled
for use in an MST topology. An example of this, where X tries to
change the DPMS property of an MST connector that was just detached from
the system:

[ 2908.320131] [drm:drm_helper_probe_single_connector_modes [drm_kms_helper]] [CONNECTOR:82:DP-6]
[ 2908.320148] [drm:drm_helper_probe_single_connector_modes [drm_kms_helper]] [CONNECTOR:82:DP-6] status updated from connected to disconnected
[ 2908.320166] [drm:drm_helper_probe_single_connector_modes [drm_kms_helper]] [CONNECTOR:82:DP-6] disconnected
[ 2908.320193] [drm:drm_mode_object_put.part.2 [drm]] OBJ ID: 111 (1)
[ 2908.320230] [drm:drm_sysfs_hotplug_event [drm]] generating hotplug event
...
[ 2908.638539] [drm:drm_ioctl [drm]] pid=12928, dev=0xe201, auth=1, DRM_IOCTL_MODE_SETPROPERTY
[ 2908.638546] [drm:drm_atomic_state_init [drm]] Allocated atomic state 000000007155ba49
[ 2908.638553] [drm:drm_mode_object_get [drm]] OBJ ID: 114 (1)
[ 2908.638560] [drm:drm_mode_object_get [drm]] OBJ ID: 108 (1)
[ 2908.638568] [drm:drm_atomic_get_crtc_state [drm]] Added [CRTC:41:head-0] 0000000097a6396e state to 000000007155ba49
[ 2908.638575] [drm:drm_atomic_add_affected_connectors [drm]] Adding all current connectors for [CRTC:41:head-0] to 000000007155ba49
[ 2908.638582] [drm:drm_mode_object_get [drm]] OBJ ID: 82 (3)
[ 2908.638589] [drm:drm_mode_object_get [drm]] OBJ ID: 82 (4)
[ 2908.638596] [drm:drm_atomic_get_connector_state [drm]] Added [CONNECTOR:82:DP-6] 0000000087427144 state to 000000007155ba49
[ 2908.638603] [drm:drm_atomic_check_only [drm]] checking 000000007155ba49
[ 2908.638609] [drm:drm_atomic_helper_check_modeset [drm_kms_helper]] [CRTC:41:head-0] active changed
[ 2908.638613] [drm:drm_atomic_helper_check_modeset [drm_kms_helper]] Updating routing for [CONNECTOR:82:DP-6]
[ 2908.638616] [drm:drm_atomic_helper_check_modeset [drm_kms_helper]] No suitable encoder found for [CONNECTOR:82:DP-6]
[ 2908.638623] [drm:drm_atomic_check_only [drm]] atomic driver check for 000000007155ba49 failed: -22
[ 2908.638630] [drm:drm_atomic_state_default_clear [drm]] Clearing atomic state 000000007155ba49
[ 2908.638637] [drm:drm_mode_object_put.part.2 [drm]] OBJ ID: 82 (4)
[ 2908.638643] [drm:drm_mode_object_put.part.2 [drm]] OBJ ID: 82 (3)
[ 2908.638650] [drm:drm_mode_object_put.part.2 [drm]] OBJ ID: 114 (2)
[ 2908.638656] [drm:drm_mode_object_put.part.2 [drm]] OBJ ID: 108 (2)
[ 2908.638663] [drm:__drm_atomic_state_free [drm]] Freeing atomic state 000000007155ba49
[ 2908.638669] [drm:drm_mode_object_put.part.2 [drm]] OBJ ID: 82 (2)
[ 2908.638676] [drm:drm_ioctl [drm]] pid=12928, ret = -22

While this doesn't usually result in any errors that would be obvious to
the user, it does result in us leaving display resources on. This in
turn leads to unwanted sideaffects like inactive GPUs being left on
(usually from the resulting leaked runtime PM ref).

So, provide an easier way of doing this that doesn't require breaking
->best_encoder(): add a common drm_dp_mst_connector_atomic_check()
function that DRM drivers can call in order to have CRTC enabling
commits fail automatically if the MST port driving the connector no
longer exists. We'll also be able to expand upon this later as well once
we add MST fallback retraining support.

Changes since v1:
- Use list_for_each_entry_safe in drm_dp_mst_connector_still_exists() -
  Julia Lawall

Signed-off-by: Lyude Paul <lyude@redhat.com>
Cc: Julia Lawall <julia.lawall@lip6.fr>
Cc: stable@vger.kernel.org
---
 drivers/gpu/drm/drm_dp_mst_topology.c | 76 +++++++++++++++++++++++++++
 include/drm/drm_dp_mst_helper.h       |  3 ++
 2 files changed, 79 insertions(+)

diff --git a/drivers/gpu/drm/drm_dp_mst_topology.c b/drivers/gpu/drm/drm_dp_mst_topology.c
index 7780567aa669..58b9554711c7 100644
--- a/drivers/gpu/drm/drm_dp_mst_topology.c
+++ b/drivers/gpu/drm/drm_dp_mst_topology.c
@@ -3129,6 +3129,82 @@ static const struct drm_private_state_funcs mst_state_funcs = {
 	.atomic_destroy_state = drm_dp_mst_destroy_state,
 };
 
+static bool
+drm_dp_mst_connector_still_exists(struct drm_connector *connector,
+				  struct drm_dp_mst_topology_mgr *mgr,
+				  struct drm_dp_mst_branch *mstb)
+{
+	struct drm_dp_mst_port *port, *tmp;
+	bool exists = false;
+
+	mstb = drm_dp_get_validated_mstb_ref(mgr, mstb);
+	if (!mstb)
+		return false;
+
+	list_for_each_entry_safe(port, tmp, &mstb->ports, next) {
+		port = drm_dp_get_validated_port_ref(mgr, port);
+		if (!port)
+			continue;
+
+		exists = (port->connector == connector ||
+			  (port->mstb &&
+			   drm_dp_mst_connector_still_exists(connector, mgr,
+							     port->mstb)));
+
+		drm_dp_put_port(port);
+		if (exists)
+			break;
+	}
+
+	drm_dp_put_mst_branch_device(mstb);
+	return exists;
+}
+
+/**
+ * drm_dp_mst_connector_atomic_check - Helper for validating a new atomic
+ *                                     state on an MST connector
+ * @connector: drm connector
+ * @connector_state: the new atomic state of @connector
+ * @mgr: the MST topology mgr for @connector
+ *
+ * This function performs various atomic checks that apply to all drivers
+ * using the DRM DP MST helpers. This should be called by all drivers at the
+ * start of the atomic_check function for their MST connectors.
+ *
+ * Return 0 for success, or negative error code on failure.
+ */
+int
+drm_dp_mst_connector_atomic_check(struct drm_connector *connector,
+				  struct drm_connector_state *connector_state,
+				  struct drm_dp_mst_topology_mgr *mgr)
+{
+	struct drm_atomic_state *state = connector_state->state;
+	struct drm_crtc *crtc = connector_state->crtc;
+	struct drm_crtc_state *new_crtc_state;
+
+	if (!crtc)
+		return 0;
+
+	new_crtc_state = drm_atomic_get_new_crtc_state(state, crtc);
+	if (!new_crtc_state)
+		return 0;
+
+	if (!drm_atomic_crtc_needs_modeset(new_crtc_state) ||
+	    !new_crtc_state->active)
+		return 0;
+
+	/* Make sure that the port for this MST connector still exists */
+	if (!drm_dp_mst_connector_still_exists(connector, mgr,
+					       mgr->mst_primary)) {
+		DRM_DEBUG_ATOMIC("[CONNECTOR:%d:%s] has disappeared from the MST topology\n",
+				 connector->base.id, connector->name);
+		return -EINVAL;
+	}
+
+	return 0;
+}
+EXPORT_SYMBOL(drm_dp_mst_connector_atomic_check);
+
 /**
  * drm_atomic_get_mst_topology_state: get MST topology state
  *
diff --git a/include/drm/drm_dp_mst_helper.h b/include/drm/drm_dp_mst_helper.h
index 7f78d26a0766..8e33c2c85d1e 100644
--- a/include/drm/drm_dp_mst_helper.h
+++ b/include/drm/drm_dp_mst_helper.h
@@ -625,6 +625,9 @@ void drm_dp_mst_topology_mgr_suspend(struct drm_dp_mst_topology_mgr *mgr);
 int drm_dp_mst_topology_mgr_resume(struct drm_dp_mst_topology_mgr *mgr);
 struct drm_dp_mst_topology_state *drm_atomic_get_mst_topology_state(struct drm_atomic_state *state,
 								    struct drm_dp_mst_topology_mgr *mgr);
+int drm_dp_mst_connector_atomic_check(struct drm_connector *connector,
+				      struct drm_connector_state *connector_state,
+				      struct drm_dp_mst_topology_mgr *mgr);
 int drm_dp_atomic_find_vcpi_slots(struct drm_atomic_state *state,
 				  struct drm_dp_mst_topology_mgr *mgr,
 				  struct drm_dp_mst_port *port, int pbn);
-- 
2.17.1

Re: [PATCH v2 1/6] drm/dp_mst: Introduce drm_dp_mst_connector_atomic_check()

[Harry Wentland]

On 2018-09-19 07:08 PM, Lyude Paul wrote:
> Currently the way that we prevent userspace from performing new modesets
> on MST connectors that have just been destroyed is rather broken.
> There's nothing in the actual DRM DP MST topology helpers that checks
> whether or not a connector still exists, instead each DRM driver does
> this on it's own, usually by returning NULL from the best_encoder
> callback which in turn, causes the atomic commit to fail.
> 
> However, this is wrong in a rather subtle way. If ->best_encoder()
> returns NULL, this makes ALL modesets involving the connector fail. This
> includes modesets from userspace that would shut off the CRTCs being
> used by the connector. Since this results in blocking any changes to a
> connector's DPMS prop, it has the sideaffect of preventing legacy
> modesetting users from ever disabling a CRTC that was previously enabled
> for use in an MST topology. An example of this, where X tries to
> change the DPMS property of an MST connector that was just detached from
> the system:
> 
> [ 2908.320131] [drm:drm_helper_probe_single_connector_modes [drm_kms_helper]] [CONNECTOR:82:DP-6]
> [ 2908.320148] [drm:drm_helper_probe_single_connector_modes [drm_kms_helper]] [CONNECTOR:82:DP-6] status updated from connected to disconnected
> [ 2908.320166] [drm:drm_helper_probe_single_connector_modes [drm_kms_helper]] [CONNECTOR:82:DP-6] disconnected
> [ 2908.320193] [drm:drm_mode_object_put.part.2 [drm]] OBJ ID: 111 (1)
> [ 2908.320230] [drm:drm_sysfs_hotplug_event [drm]] generating hotplug event
> ...
> [ 2908.638539] [drm:drm_ioctl [drm]] pid=12928, dev=0xe201, auth=1, DRM_IOCTL_MODE_SETPROPERTY
> [ 2908.638546] [drm:drm_atomic_state_init [drm]] Allocated atomic state 000000007155ba49
> [ 2908.638553] [drm:drm_mode_object_get [drm]] OBJ ID: 114 (1)
> [ 2908.638560] [drm:drm_mode_object_get [drm]] OBJ ID: 108 (1)
> [ 2908.638568] [drm:drm_atomic_get_crtc_state [drm]] Added [CRTC:41:head-0] 0000000097a6396e state to 000000007155ba49
> [ 2908.638575] [drm:drm_atomic_add_affected_connectors [drm]] Adding all current connectors for [CRTC:41:head-0] to 000000007155ba49
> [ 2908.638582] [drm:drm_mode_object_get [drm]] OBJ ID: 82 (3)
> [ 2908.638589] [drm:drm_mode_object_get [drm]] OBJ ID: 82 (4)
> [ 2908.638596] [drm:drm_atomic_get_connector_state [drm]] Added [CONNECTOR:82:DP-6] 0000000087427144 state to 000000007155ba49
> [ 2908.638603] [drm:drm_atomic_check_only [drm]] checking 000000007155ba49
> [ 2908.638609] [drm:drm_atomic_helper_check_modeset [drm_kms_helper]] [CRTC:41:head-0] active changed
> [ 2908.638613] [drm:drm_atomic_helper_check_modeset [drm_kms_helper]] Updating routing for [CONNECTOR:82:DP-6]
> [ 2908.638616] [drm:drm_atomic_helper_check_modeset [drm_kms_helper]] No suitable encoder found for [CONNECTOR:82:DP-6]
> [ 2908.638623] [drm:drm_atomic_check_only [drm]] atomic driver check for 000000007155ba49 failed: -22
> [ 2908.638630] [drm:drm_atomic_state_default_clear [drm]] Clearing atomic state 000000007155ba49
> [ 2908.638637] [drm:drm_mode_object_put.part.2 [drm]] OBJ ID: 82 (4)
> [ 2908.638643] [drm:drm_mode_object_put.part.2 [drm]] OBJ ID: 82 (3)
> [ 2908.638650] [drm:drm_mode_object_put.part.2 [drm]] OBJ ID: 114 (2)
> [ 2908.638656] [drm:drm_mode_object_put.part.2 [drm]] OBJ ID: 108 (2)
> [ 2908.638663] [drm:__drm_atomic_state_free [drm]] Freeing atomic state 000000007155ba49
> [ 2908.638669] [drm:drm_mode_object_put.part.2 [drm]] OBJ ID: 82 (2)
> [ 2908.638676] [drm:drm_ioctl [drm]] pid=12928, ret = -22
> 
> While this doesn't usually result in any errors that would be obvious to
> the user, it does result in us leaving display resources on. This in
> turn leads to unwanted sideaffects like inactive GPUs being left on
> (usually from the resulting leaked runtime PM ref).
> 
> So, provide an easier way of doing this that doesn't require breaking
> ->best_encoder(): add a common drm_dp_mst_connector_atomic_check()
> function that DRM drivers can call in order to have CRTC enabling
> commits fail automatically if the MST port driving the connector no
> longer exists. We'll also be able to expand upon this later as well once
> we add MST fallback retraining support.
> 
> Changes since v1:
> - Use list_for_each_entry_safe in drm_dp_mst_connector_still_exists() -
>   Julia Lawall
> 
> Signed-off-by: Lyude Paul <lyude@redhat.com>
> Cc: Julia Lawall <julia.lawall@lip6.fr>
> Cc: stable@vger.kernel.org

Whoops, missed the v2 earlier. It's still
Acked-by: Harry Wentland <harry.wentland@amd.com>

Harry

> ---
>  drivers/gpu/drm/drm_dp_mst_topology.c | 76 +++++++++++++++++++++++++++
>  include/drm/drm_dp_mst_helper.h       |  3 ++
>  2 files changed, 79 insertions(+)
> 
> diff --git a/drivers/gpu/drm/drm_dp_mst_topology.c b/drivers/gpu/drm/drm_dp_mst_topology.c
> index 7780567aa669..58b9554711c7 100644
> --- a/drivers/gpu/drm/drm_dp_mst_topology.c
> +++ b/drivers/gpu/drm/drm_dp_mst_topology.c
> @@ -3129,6 +3129,82 @@ static const struct drm_private_state_funcs mst_state_funcs = {
>  	.atomic_destroy_state = drm_dp_mst_destroy_state,
>  };
>  
> +static bool
> +drm_dp_mst_connector_still_exists(struct drm_connector *connector,
> +				  struct drm_dp_mst_topology_mgr *mgr,
> +				  struct drm_dp_mst_branch *mstb)
> +{
> +	struct drm_dp_mst_port *port, *tmp;
> +	bool exists = false;
> +
> +	mstb = drm_dp_get_validated_mstb_ref(mgr, mstb);
> +	if (!mstb)
> +		return false;
> +
> +	list_for_each_entry_safe(port, tmp, &mstb->ports, next) {
> +		port = drm_dp_get_validated_port_ref(mgr, port);
> +		if (!port)
> +			continue;
> +
> +		exists = (port->connector == connector ||
> +			  (port->mstb &&
> +			   drm_dp_mst_connector_still_exists(connector, mgr,
> +							     port->mstb)));
> +
> +		drm_dp_put_port(port);
> +		if (exists)
> +			break;
> +	}
> +
> +	drm_dp_put_mst_branch_device(mstb);
> +	return exists;
> +}
> +
> +/**
> + * drm_dp_mst_connector_atomic_check - Helper for validating a new atomic
> + *                                     state on an MST connector
> + * @connector: drm connector
> + * @connector_state: the new atomic state of @connector
> + * @mgr: the MST topology mgr for @connector
> + *
> + * This function performs various atomic checks that apply to all drivers
> + * using the DRM DP MST helpers. This should be called by all drivers at the
> + * start of the atomic_check function for their MST connectors.
> + *
> + * Return 0 for success, or negative error code on failure.
> + */
> +int
> +drm_dp_mst_connector_atomic_check(struct drm_connector *connector,
> +				  struct drm_connector_state *connector_state,
> +				  struct drm_dp_mst_topology_mgr *mgr)
> +{
> +	struct drm_atomic_state *state = connector_state->state;
> +	struct drm_crtc *crtc = connector_state->crtc;
> +	struct drm_crtc_state *new_crtc_state;
> +
> +	if (!crtc)
> +		return 0;
> +
> +	new_crtc_state = drm_atomic_get_new_crtc_state(state, crtc);
> +	if (!new_crtc_state)
> +		return 0;
> +
> +	if (!drm_atomic_crtc_needs_modeset(new_crtc_state) ||
> +	    !new_crtc_state->active)
> +		return 0;
> +
> +	/* Make sure that the port for this MST connector still exists */
> +	if (!drm_dp_mst_connector_still_exists(connector, mgr,
> +					       mgr->mst_primary)) {
> +		DRM_DEBUG_ATOMIC("[CONNECTOR:%d:%s] has disappeared from the MST topology\n",
> +				 connector->base.id, connector->name);
> +		return -EINVAL;
> +	}
> +
> +	return 0;
> +}
> +EXPORT_SYMBOL(drm_dp_mst_connector_atomic_check);
> +
>  /**
>   * drm_atomic_get_mst_topology_state: get MST topology state
>   *
> diff --git a/include/drm/drm_dp_mst_helper.h b/include/drm/drm_dp_mst_helper.h
> index 7f78d26a0766..8e33c2c85d1e 100644
> --- a/include/drm/drm_dp_mst_helper.h
> +++ b/include/drm/drm_dp_mst_helper.h
> @@ -625,6 +625,9 @@ void drm_dp_mst_topology_mgr_suspend(struct drm_dp_mst_topology_mgr *mgr);
>  int drm_dp_mst_topology_mgr_resume(struct drm_dp_mst_topology_mgr *mgr);
>  struct drm_dp_mst_topology_state *drm_atomic_get_mst_topology_state(struct drm_atomic_state *state,
>  								    struct drm_dp_mst_topology_mgr *mgr);
> +int drm_dp_mst_connector_atomic_check(struct drm_connector *connector,
> +				      struct drm_connector_state *connector_state,
> +				      struct drm_dp_mst_topology_mgr *mgr);
>  int drm_dp_atomic_find_vcpi_slots(struct drm_atomic_state *state,
>  				  struct drm_dp_mst_topology_mgr *mgr,
>  				  struct drm_dp_mst_port *port, int pbn);
> 

[PATCH v2 2/6] drm/nouveau: Unbreak nv50_mstc->best_encoder()

[Lyude Paul]

As mentioned in the previous commit, we currently prevent new modesets
on recently-removed MST connectors by returning no encoder from our
->best_encoder() callback once the MST port has disappeared. This is
wrong however, because it prevents legacy modesetting users from being
able to disable CRTCs on MST connectors after the connector's respective
topology has disappeared.

So, fix this by instead using the new
drm_dp_mst_connector_atomic_check() helper instead while always
returning a valid encoder from ->best_encoder().

Signed-off-by: Lyude Paul <lyude@redhat.com>
Cc: stable@vger.kernel.org
---
 drivers/gpu/drm/nouveau/dispnv50/disp.c | 25 +++++++++++++++----------
 1 file changed, 15 insertions(+), 10 deletions(-)

diff --git a/drivers/gpu/drm/nouveau/dispnv50/disp.c b/drivers/gpu/drm/nouveau/dispnv50/disp.c
index 9da0bdfe1e1c..8d6f6ee9bc75 100644
--- a/drivers/gpu/drm/nouveau/dispnv50/disp.c
+++ b/drivers/gpu/drm/nouveau/dispnv50/disp.c
@@ -881,22 +881,16 @@ nv50_mstc_atomic_best_encoder(struct drm_connector *connector,
 {
 	struct nv50_head *head = nv50_head(connector_state->crtc);
 	struct nv50_mstc *mstc = nv50_mstc(connector);
-	if (mstc->port) {
-		struct nv50_mstm *mstm = mstc->mstm;
-		return &mstm->msto[head->base.index]->encoder;
-	}
-	return NULL;
+
+	return &mstc->mstm->msto[head->base.index]->encoder;
 }
 
 static struct drm_encoder *
 nv50_mstc_best_encoder(struct drm_connector *connector)
 {
 	struct nv50_mstc *mstc = nv50_mstc(connector);
-	if (mstc->port) {
-		struct nv50_mstm *mstm = mstc->mstm;
-		return &mstm->msto[0]->encoder;
-	}
-	return NULL;
+
+	return &mstc->mstm->msto[0]->encoder;
 }
 
 static enum drm_mode_status
@@ -926,10 +920,21 @@ nv50_mstc_get_modes(struct drm_connector *connector)
 	return ret;
 }
 
+static int
+nv50_mstc_atomic_check(struct drm_connector *connector,
+		       struct drm_connector_state *state)
+{
+	struct nv50_mstc *mstc = nv50_mstc(connector);
+
+	return drm_dp_mst_connector_atomic_check(connector, state,
+						 &mstc->mstm->mgr);
+}
+
 static const struct drm_connector_helper_funcs
 nv50_mstc_help = {
 	.get_modes = nv50_mstc_get_modes,
 	.mode_valid = nv50_mstc_mode_valid,
+	.atomic_check = nv50_mstc_atomic_check,
 	.best_encoder = nv50_mstc_best_encoder,
 	.atomic_best_encoder = nv50_mstc_atomic_best_encoder,
 };
-- 
2.17.1

[PATCH v2 3/6] drm/i915: Leave intel_conn->mst_port set, use mst_port_gone instead

[Lyude Paul]

Currently we set intel_connector->mst_port to NULL to signify that the
MST port has been removed from the system so that we can prevent further
action on the port such as connector probes, mode probing, etc.
However, we're going to need access to intel_connector->mst_port in
order to fixup ->best_encoder() so that it can always return the correct
encoder for an MST port to prevent legacy DPMS prop changes from
failing. This should be safe, so instead keep intel_connector->mst_port
always set and instead add intel_connector->mst_port_gone in order to
signify whether or not the connector has disappeared from the system.

Signed-off-by: Lyude Paul <lyude@redhat.com>
Cc: stable@vger.kernel.org
---
 drivers/gpu/drm/i915/intel_dp_mst.c | 14 +++++++-------
 drivers/gpu/drm/i915/intel_drv.h    |  1 +
 2 files changed, 8 insertions(+), 7 deletions(-)

diff --git a/drivers/gpu/drm/i915/intel_dp_mst.c b/drivers/gpu/drm/i915/intel_dp_mst.c
index 4ecd65375603..fcb9b87b9339 100644
--- a/drivers/gpu/drm/i915/intel_dp_mst.c
+++ b/drivers/gpu/drm/i915/intel_dp_mst.c
@@ -311,9 +311,8 @@ static int intel_dp_mst_get_ddc_modes(struct drm_connector *connector)
 	struct edid *edid;
 	int ret;
 
-	if (!intel_dp) {
+	if (intel_connector->mst_port_gone)
 		return intel_connector_update_modes(connector, NULL);
-	}
 
 	edid = drm_dp_mst_get_edid(connector, &intel_dp->mst_mgr, intel_connector->port);
 	ret = intel_connector_update_modes(connector, edid);
@@ -328,9 +327,10 @@ intel_dp_mst_detect(struct drm_connector *connector, bool force)
 	struct intel_connector *intel_connector = to_intel_connector(connector);
 	struct intel_dp *intel_dp = intel_connector->mst_port;
 
-	if (!intel_dp)
+	if (intel_connector->mst_port_gone)
 		return connector_status_disconnected;
-	return drm_dp_mst_detect_port(connector, &intel_dp->mst_mgr, intel_connector->port);
+	return drm_dp_mst_detect_port(connector, &intel_dp->mst_mgr,
+				      intel_connector->port);
 }
 
 static void
@@ -370,7 +370,7 @@ intel_dp_mst_mode_valid(struct drm_connector *connector,
 	int bpp = 24; /* MST uses fixed bpp */
 	int max_rate, mode_rate, max_lanes, max_link_clock;
 
-	if (!intel_dp)
+	if (intel_connector->mst_port_gone)
 		return MODE_ERROR;
 
 	if (mode->flags & DRM_MODE_FLAG_DBLSCAN)
@@ -402,7 +402,7 @@ static struct drm_encoder *intel_mst_atomic_best_encoder(struct drm_connector *c
 	struct intel_dp *intel_dp = intel_connector->mst_port;
 	struct intel_crtc *crtc = to_intel_crtc(state->crtc);
 
-	if (!intel_dp)
+	if (intel_connector->mst_port_gone)
 		return NULL;
 	return &intel_dp->mst_encoders[crtc->pipe]->base.base;
 }
@@ -514,7 +514,7 @@ static void intel_dp_destroy_mst_connector(struct drm_dp_mst_topology_mgr *mgr,
 						   connector);
 	/* prevent race with the check in ->detect */
 	drm_modeset_lock(&connector->dev->mode_config.connection_mutex, NULL);
-	intel_connector->mst_port = NULL;
+	intel_connector->mst_port_gone = true;
 	drm_modeset_unlock(&connector->dev->mode_config.connection_mutex);
 
 	drm_connector_put(connector);
diff --git a/drivers/gpu/drm/i915/intel_drv.h b/drivers/gpu/drm/i915/intel_drv.h
index 8fc61e96754f..87ce772ae7f8 100644
--- a/drivers/gpu/drm/i915/intel_drv.h
+++ b/drivers/gpu/drm/i915/intel_drv.h
@@ -409,6 +409,7 @@ struct intel_connector {
 	void *port; /* store this opaque as its illegal to dereference it */
 
 	struct intel_dp *mst_port;
+	bool mst_port_gone;
 
 	/* Work struct to schedule a uevent on link train failure */
 	struct work_struct modeset_retry_work;
-- 
2.17.1

Re: [PATCH v2 3/6] drm/i915: Leave intel_conn->mst_port set, use mst_port_gone instead

[Jani Nikula]

On Wed, 19 Sep 2018, Lyude Paul <lyude@redhat.com> wrote:
> Currently we set intel_connector->mst_port to NULL to signify that the
> MST port has been removed from the system so that we can prevent further
> action on the port such as connector probes, mode probing, etc.
> However, we're going to need access to intel_connector->mst_port in
> order to fixup ->best_encoder() so that it can always return the correct
> encoder for an MST port to prevent legacy DPMS prop changes from
> failing. This should be safe, so instead keep intel_connector->mst_port
> always set and instead add intel_connector->mst_port_gone in order to
> signify whether or not the connector has disappeared from the system.
>
> Signed-off-by: Lyude Paul <lyude@redhat.com>
> Cc: stable@vger.kernel.org
> ---
>  drivers/gpu/drm/i915/intel_dp_mst.c | 14 +++++++-------
>  drivers/gpu/drm/i915/intel_drv.h    |  1 +
>  2 files changed, 8 insertions(+), 7 deletions(-)
>
> diff --git a/drivers/gpu/drm/i915/intel_dp_mst.c b/drivers/gpu/drm/i915/intel_dp_mst.c
> index 4ecd65375603..fcb9b87b9339 100644
> --- a/drivers/gpu/drm/i915/intel_dp_mst.c
> +++ b/drivers/gpu/drm/i915/intel_dp_mst.c
> @@ -311,9 +311,8 @@ static int intel_dp_mst_get_ddc_modes(struct drm_connector *connector)
>  	struct edid *edid;
>  	int ret;
>  
> -	if (!intel_dp) {
> +	if (intel_connector->mst_port_gone)
>  		return intel_connector_update_modes(connector, NULL);
> -	}
>  
>  	edid = drm_dp_mst_get_edid(connector, &intel_dp->mst_mgr, intel_connector->port);
>  	ret = intel_connector_update_modes(connector, edid);
> @@ -328,9 +327,10 @@ intel_dp_mst_detect(struct drm_connector *connector, bool force)
>  	struct intel_connector *intel_connector = to_intel_connector(connector);
>  	struct intel_dp *intel_dp = intel_connector->mst_port;
>  
> -	if (!intel_dp)
> +	if (intel_connector->mst_port_gone)
>  		return connector_status_disconnected;
> -	return drm_dp_mst_detect_port(connector, &intel_dp->mst_mgr, intel_connector->port);
> +	return drm_dp_mst_detect_port(connector, &intel_dp->mst_mgr,
> +				      intel_connector->port);
>  }
>  
>  static void
> @@ -370,7 +370,7 @@ intel_dp_mst_mode_valid(struct drm_connector *connector,
>  	int bpp = 24; /* MST uses fixed bpp */
>  	int max_rate, mode_rate, max_lanes, max_link_clock;
>  
> -	if (!intel_dp)
> +	if (intel_connector->mst_port_gone)
>  		return MODE_ERROR;
>  
>  	if (mode->flags & DRM_MODE_FLAG_DBLSCAN)
> @@ -402,7 +402,7 @@ static struct drm_encoder *intel_mst_atomic_best_encoder(struct drm_connector *c
>  	struct intel_dp *intel_dp = intel_connector->mst_port;
>  	struct intel_crtc *crtc = to_intel_crtc(state->crtc);
>  
> -	if (!intel_dp)
> +	if (intel_connector->mst_port_gone)
>  		return NULL;
>  	return &intel_dp->mst_encoders[crtc->pipe]->base.base;
>  }
> @@ -514,7 +514,7 @@ static void intel_dp_destroy_mst_connector(struct drm_dp_mst_topology_mgr *mgr,
>  						   connector);
>  	/* prevent race with the check in ->detect */
>  	drm_modeset_lock(&connector->dev->mode_config.connection_mutex, NULL);
> -	intel_connector->mst_port = NULL;
> +	intel_connector->mst_port_gone = true;
>  	drm_modeset_unlock(&connector->dev->mode_config.connection_mutex);
>  
>  	drm_connector_put(connector);
> diff --git a/drivers/gpu/drm/i915/intel_drv.h b/drivers/gpu/drm/i915/intel_drv.h
> index 8fc61e96754f..87ce772ae7f8 100644
> --- a/drivers/gpu/drm/i915/intel_drv.h
> +++ b/drivers/gpu/drm/i915/intel_drv.h
> @@ -409,6 +409,7 @@ struct intel_connector {
>  	void *port; /* store this opaque as its illegal to dereference it */
>  
>  	struct intel_dp *mst_port;

	/*
	 * please explain mst_port_gone here too instead of just the
	 * commit message!
	 */

> +	bool mst_port_gone;
>  
>  	/* Work struct to schedule a uevent on link train failure */
>  	struct work_struct modeset_retry_work;

-- 
Jani Nikula, Intel Open Source Graphics Center

[PATCH v2 4/6] drm/i915: Skip vcpi allocation for MSTB ports that are gone

[Lyude Paul]

Since we need to be able to allow DPMS on->off prop changes after an MST
port has disappeared from the system, we need to be able to make sure we
can compute a config for the resulting atomic commit. Currently this is
impossible when the port has disappeared, since the VCPI slot searching
we try to do in intel_dp_mst_compute_config() will fail with -EINVAL.

Since the only commits we want to allow on no-longer-present MST ports
are ones that shut off display hardware, we already know that no VCPI
allocations are needed. So, hardcode the VCPI slot count to 0 when
intel_dp_mst_compute_config() is called on an MST port that's gone.

Signed-off-by: Lyude Paul <lyude@redhat.com>
Cc: stable@vger.kernel.org
---
 drivers/gpu/drm/i915/intel_dp_mst.c | 17 +++++++++++------
 1 file changed, 11 insertions(+), 6 deletions(-)

diff --git a/drivers/gpu/drm/i915/intel_dp_mst.c b/drivers/gpu/drm/i915/intel_dp_mst.c
index fcb9b87b9339..a366f32b048a 100644
--- a/drivers/gpu/drm/i915/intel_dp_mst.c
+++ b/drivers/gpu/drm/i915/intel_dp_mst.c
@@ -42,7 +42,7 @@ static bool intel_dp_mst_compute_config(struct intel_encoder *encoder,
 		to_intel_connector(conn_state->connector);
 	struct drm_atomic_state *state = pipe_config->base.state;
 	int bpp;
-	int lane_count, slots;
+	int lane_count, slots = 0;
 	const struct drm_display_mode *adjusted_mode = &pipe_config->base.adjusted_mode;
 	int mst_pbn;
 	bool reduce_m_n = drm_dp_has_quirk(&intel_dp->desc,
@@ -76,11 +76,16 @@ static bool intel_dp_mst_compute_config(struct intel_encoder *encoder,
 	mst_pbn = drm_dp_calc_pbn_mode(adjusted_mode->crtc_clock, bpp);
 	pipe_config->pbn = mst_pbn;
 
-	slots = drm_dp_atomic_find_vcpi_slots(state, &intel_dp->mst_mgr,
-					      connector->port, mst_pbn);
-	if (slots < 0) {
-		DRM_DEBUG_KMS("failed finding vcpi slots:%d\n", slots);
-		return false;
+	if (!connector->mst_port_gone) {
+		slots = drm_dp_atomic_find_vcpi_slots(state,
+						      &intel_dp->mst_mgr,
+						      connector->port,
+						      mst_pbn);
+		if (slots < 0) {
+			DRM_DEBUG_KMS("failed finding vcpi slots:%d\n",
+				      slots);
+			return false;
+		}
 	}
 
 	intel_link_compute_m_n(bpp, lane_count,
-- 
2.17.1

[PATCH v2 5/6] drm/i915: Fix intel_dp_mst_best_encoder()

[Lyude Paul]

Currently, i915 appears to rely on blocking modesets on
no-longer-present MSTB ports by simply returning NULL for
->best_encoder(), which in turn causes any new atomic commits that don't
disable the CRTC to fail. This is wrong however, since we still want to
allow userspace to disable CRTCs on no-longer-present MSTB ports by
changing the DPMS state to off and this still requires that we retrieve
an encoder.

So, fix this by always returning a valid encoder regardless of the state
of the MST port. Additionally, make intel_dp_mst_atomic_check() simply
rely on drm_dp_mst_connector_atomic_check() to prevent new modesets on
no-longer-present MSTB ports.

Signed-off-by: Lyude Paul <lyude@redhat.com>
Cc: stable@vger.kernel.org
---
 drivers/gpu/drm/i915/intel_dp_mst.c | 17 ++++++++---------
 1 file changed, 8 insertions(+), 9 deletions(-)

diff --git a/drivers/gpu/drm/i915/intel_dp_mst.c b/drivers/gpu/drm/i915/intel_dp_mst.c
index a366f32b048a..2b798d4592f0 100644
--- a/drivers/gpu/drm/i915/intel_dp_mst.c
+++ b/drivers/gpu/drm/i915/intel_dp_mst.c
@@ -106,14 +106,21 @@ static bool intel_dp_mst_compute_config(struct intel_encoder *encoder,
 }
 
 static int intel_dp_mst_atomic_check(struct drm_connector *connector,
-		struct drm_connector_state *new_conn_state)
+				     struct drm_connector_state *new_conn_state)
 {
 	struct drm_atomic_state *state = new_conn_state->state;
 	struct drm_connector_state *old_conn_state;
 	struct drm_crtc *old_crtc;
 	struct drm_crtc_state *crtc_state;
+	struct drm_dp_mst_topology_mgr *mgr =
+		&to_intel_connector(connector)->mst_port->mst_mgr;
 	int slots, ret = 0;
 
+	ret = drm_dp_mst_connector_atomic_check(connector, new_conn_state,
+						mgr);
+	if (ret)
+		return ret;
+
 	old_conn_state = drm_atomic_get_old_connector_state(state, connector);
 	old_crtc = old_conn_state->crtc;
 	if (!old_crtc)
@@ -122,12 +129,6 @@ static int intel_dp_mst_atomic_check(struct drm_connector *connector,
 	crtc_state = drm_atomic_get_new_crtc_state(state, old_crtc);
 	slots = to_intel_crtc_state(crtc_state)->dp_m_n.tu;
 	if (drm_atomic_crtc_needs_modeset(crtc_state) && slots > 0) {
-		struct drm_dp_mst_topology_mgr *mgr;
-		struct drm_encoder *old_encoder;
-
-		old_encoder = old_conn_state->best_encoder;
-		mgr = &enc_to_mst(old_encoder)->primary->dp.mst_mgr;
-
 		ret = drm_dp_atomic_release_vcpi_slots(state, mgr, slots);
 		if (ret)
 			DRM_DEBUG_KMS("failed releasing %d vcpi slots:%d\n", slots, ret);
@@ -407,8 +408,6 @@ static struct drm_encoder *intel_mst_atomic_best_encoder(struct drm_connector *c
 	struct intel_dp *intel_dp = intel_connector->mst_port;
 	struct intel_crtc *crtc = to_intel_crtc(state->crtc);
 
-	if (intel_connector->mst_port_gone)
-		return NULL;
 	return &intel_dp->mst_encoders[crtc->pipe]->base.base;
 }
 
-- 
2.17.1

[PATCH v2 6/6] drm/amdgpu/dm/mst: Use drm_dp_mst_connector_atomic_check()

[Lyude Paul]

Hook this into amdgpu's atomic check for their connectors so they never
get modesets on no-longer-present MST connectors. We'll also expand on
this later once we add DP MST fallback retraining support.

As well, turns out that the only atomic DRM driver without the
->best_encoder() bug is amdgpu. Congrats AMD!

Signed-off-by: Lyude Paul <lyude@redhat.com>
---
 .../drm/amd/display/amdgpu_dm/amdgpu_dm_mst_types.c  | 12 ++++++++++++
 1 file changed, 12 insertions(+)

diff --git a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_mst_types.c b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_mst_types.c
index 9a300732ba37..d011a39f17b2 100644
--- a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_mst_types.c
+++ b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_mst_types.c
@@ -294,10 +294,22 @@ static struct drm_encoder *dm_mst_best_encoder(struct drm_connector *connector)
 	return &amdgpu_dm_connector->mst_encoder->base;
 }
 
+static int
+amdgpu_dm_mst_connector_atomic_check(struct drm_connector *connector,
+				     struct drm_connector_state *new_cstate)
+{
+	struct amdgpu_dm_connector *aconnector =
+		to_amdgpu_dm_connector(connector);
+
+	return drm_dp_mst_connector_atomic_check(connector, new_cstate,
+						 &aconnector->mst_mgr);
+}
+
 static const struct drm_connector_helper_funcs dm_dp_mst_connector_helper_funcs = {
 	.get_modes = dm_dp_mst_get_modes,
 	.mode_valid = amdgpu_dm_connector_mode_valid,
 	.best_encoder = dm_mst_best_encoder,
+	.atomic_check = amdgpu_dm_mst_connector_atomic_check,
 };
 
 static void amdgpu_dm_encoder_destroy(struct drm_encoder *encoder)
-- 
2.17.1