On Mon, Dec 05, 2016 at 04:27:35PM -0800, Manasi Navare wrote:

At the time userspace does setcrtc, we've already promised the mode would work. The promise is based on the theoretical capabilities of the link, but it's possible we can't reach this in practice. The DP spec describes how the link should be reduced, but we can't reduce the link below the requirements of the mode. Black screen follows.

One idea would be to have setcrtc return a failure. However, it already should not fail as the atomic checks have passed. It would also conflict with the idea of making setcrtc asynchronous in the future, returning before the actual mode setting and link training.

Another idea is to train the link "upfront" at hotplug time, before pruning the mode list, so that we can do the pruning based on practical not theoretical capabilities. However, the changes for link training are pretty drastic, all for the sake of error handling and DP compliance, when the most common happy day scenario is the current approach of link training at mode setting time, using the optimal parameters for the mode. It is also not certain all hardware could do this without the pipe on; not even all our hardware can do this. Some of this can be solved, but not trivially.

Both of the above ideas also fail to address link degradation *during* operation.

The solution is to add a new "link-status" connector property in order to address link training failure in a way that: a) changes the current happy day scenario as little as possible, to avoid regressions, b) can be implemented the same way by all drm drivers, c) is still opt-in for the drivers and userspace, and opting out doesn't regress the user experience, d) doesn't prevent drivers from implementing better or alternate approaches, possibly without userspace involvement. And, of course, handles all the issues presented. In the usual happy day scenario, this is always "good". If something fails during or after a mode set, the kernel driver can set the link status to "bad" and issue a hotplug uevent for userspace to have it re-check the valid modes through GET_CONNECTOR IOCTL, and try modeset again. If the theoretical capabilities of the link can't be reached, the mode list is trimmed based on that.

v4:

• Add comments in kernel-doc format (Daniel Vetter)
• Update the kernel-doc for link-status (Sean Paul)

v3:

• Fixed a build error (Jani Saarinen)

v2:

• Removed connector->link_status (Daniel Vetter)
• Set connector->state->link_status in drm_mode_connector_set_link_status_property

(Daniel Vetter)

• Set the connector_changed flag to true if connector->state->link_status changed.
• Reset link_status to GOOD in update_output_state (Daniel Vetter)
• Never allow userspace to set link status from Good To Bad (Daniel Vetter)

Acked-by: Tony Cheng tony.cheng@amd.com Acked-by: Harry Wentland harry.wentland@amd.com Cc: Jani Nikula jani.nikula@linux.intel.com Cc: Daniel Vetter daniel.vetter@intel.com Cc: Ville Syrjala ville.syrjala@linux.intel.com Cc: Chris Wilson chris@chris-wilson.co.uk Cc: Sean Paul seanpaul@chromium.org

Didn't Sean r-b this already?

Signed-off-by: Manasi Navare manasi.d.navare@intel.com

Looks all nice and tidy, great work.

Reviewed-by: Daniel Vetter daniel.vetter@ffwll.ch

Can be merged as soon as we have an ack from Martin that the -modesetting side is ready to go too. -Daniel

---

drivers/gpu/drm/drm_atomic.c | 10 +++++++ drivers/gpu/drm/drm_atomic_helper.c | 15 +++++++++++ drivers/gpu/drm/drm_connector.c | 53 +++++++++++++++++++++++++++++++++++++ include/drm/drm_connector.h | 19 +++++++++++++ include/drm/drm_mode_config.h | 5 ++++ include/uapi/drm/drm_mode.h | 4 +++ 6 files changed, 106 insertions(+)

diff --git a/drivers/gpu/drm/drm_atomic.c b/drivers/gpu/drm/drm_atomic.c index 19d7bcb..70a5152 100644 --- a/drivers/gpu/drm/drm_atomic.c +++ b/drivers/gpu/drm/drm_atomic.c @@ -1087,6 +1087,14 @@ int drm_atomic_connector_set_property(struct drm_connector *connector, * now?) atomic writes to DPMS property: */ return -EINVAL;

• } else if (property == config->link_status_property) {

• /* Never downgrade from GOOD to BAD on userspace's request here,
  

•  * only hw issues can do that.
  

•  */
  

• if (state->link_status == DRM_LINK_STATUS_GOOD)
  

• 	return 0;
  

• state->link_status = val;
  

• return 0;
  

  } else if (connector->funcs->atomic_set_property) { return connector->funcs->atomic_set_property(connector, state, property, val);

@@ -1135,6 +1143,8 @@ static void drm_atomic_connector_print_state(struct drm_printer *p, *val = (state->crtc) ? state->crtc->base.id : 0; } else if (property == config->dpms_property) { *val = connector->dpms;

• } else if (property == config->link_status_property) {

• *val = state->link_status;
  

  } else if (connector->funcs->atomic_get_property) { return connector->funcs->atomic_get_property(connector, state, property, val);

diff --git a/drivers/gpu/drm/drm_atomic_helper.c b/drivers/gpu/drm/drm_atomic_helper.c index 494680c..3d6c1ce 100644 --- a/drivers/gpu/drm/drm_atomic_helper.c +++ b/drivers/gpu/drm/drm_atomic_helper.c @@ -519,6 +519,13 @@ static int handle_conflicting_encoders(struct drm_atomic_state *state, connector_state); if (ret) return ret;

• if (connector->state->crtc) {
  

• 	crtc_state = drm_atomic_get_existing_crtc_state(state,
  

• 							connector->state->crtc);
  

• 	if (connector->state->link_status !=
  

• 	    connector_state->link_status)
  

• 		crtc_state->connectors_changed = true;
  

• }
  

  }

  /*

@@ -2258,6 +2265,8 @@ static int update_output_state(struct drm_atomic_state *state, NULL); if (ret) return ret;

• 	/* Make sure legacy setCrtc always re-trains */
  

• 	conn_state->link_status = DRM_LINK_STATUS_GOOD;
  

  } }

@@ -2301,6 +2310,12 @@ static int update_output_state(struct drm_atomic_state *state,

• Provides a default crtc set_config handler using the atomic driver interface.

• • NOTE: For backwards compatibility with old userspace this automatically
• • resets the "link-status" property to GOOD, to force any link
• • re-training. The SETCRTC ioctl does not define whether an update does
• • need a full modeset or just a plane update, hence we're allowed to do
• • that. See also drm_mode_connector_set_link_status_property().
• • Returns:
  • Returns 0 on success, negative errno numbers on failure.
  */

diff --git a/drivers/gpu/drm/drm_connector.c b/drivers/gpu/drm/drm_connector.c index 5a45262..eee038a 100644 --- a/drivers/gpu/drm/drm_connector.c +++ b/drivers/gpu/drm/drm_connector.c @@ -243,6 +243,10 @@ int drm_connector_init(struct drm_device *dev, drm_object_attach_property(&connector->base, config->dpms_property, 0);

• drm_object_attach_property(&connector->base,

• 		   config->link_status_property,
  

• 		   0);
  

• if (drm_core_check_feature(dev, DRIVER_ATOMIC)) { drm_object_attach_property(&connector->base, config->prop_crtc_id, 0); }

@@ -506,6 +510,12 @@ const char *drm_get_subpixel_order_name(enum subpixel_order order) }; DRM_ENUM_NAME_FN(drm_get_dpms_name, drm_dpms_enum_list)

+static const struct drm_prop_enum_list drm_link_status_enum_list[] = {

• { DRM_MODE_LINK_STATUS_GOOD, "Good" },
• { DRM_MODE_LINK_STATUS_BAD, "Bad" },

+}; +DRM_ENUM_NAME_FN(drm_get_link_status_name, drm_link_status_enum_list)

/**

• drm_display_info_set_bus_formats - set the supported bus formats
• @info: display info to store bus formats in

@@ -625,6 +635,11 @@ int drm_display_info_set_bus_formats(struct drm_display_info *info,

• tiling and virtualize both &drm_crtc and &drm_plane if needed. Drivers
• should update this value using drm_mode_connector_set_tile_property().
• Userspace cannot change this property.

• • link-status:

• •  Connector link-status property to indicate the status of link. The default
    

• •  value of link-status is "GOOD". If something fails during or after modeset,
    

• •  the kernel driver may set this to "BAD" and issue a hotplug uevent. Drivers
    

• •  should update this value using drm_mode_connector_set_link_status_property().
    

  • Connectors also have one standardized atomic property:

@@ -666,6 +681,13 @@ int drm_connector_create_standard_properties(struct drm_device *dev) return -ENOMEM; dev->mode_config.tile_property = prop;

• prop = drm_property_create_enum(dev, 0, "link-status",

• 			drm_link_status_enum_list,
  

• 			ARRAY_SIZE(drm_link_status_enum_list));
  

• if (!prop)

• return -ENOMEM;
  

• dev->mode_config.link_status_property = prop;
• return 0;

}

@@ -995,6 +1017,37 @@ int drm_mode_connector_update_edid_property(struct drm_connector *connector, } EXPORT_SYMBOL(drm_mode_connector_update_edid_property);

+/**

• • drm_mode_connector_set_link_status_property - Set link status property of a connector
• • @connector: drm connector
• • @link_status: new value of link status property (0: Good, 1: Bad)
• • In usual working scenario, this link status property will always be set to
• • "GOOD". If something fails during or after a mode set, the kernel driver
• • may set this link status property to "BAD". The caller then needs to send a
• • hotplug uevent for userspace to re-check the valid modes through
• • GET_CONNECTOR_IOCTL and retry modeset.
• • Note: Drivers cannot rely on userspace to support this property and
• • issue a modeset. As such, they may choose to handle issues (like
• • re-training a link) without userspace's intervention.
• • The reason for adding this property is to handle link training failures, but
• • it is not limited to DP or link training. For example, if we implement
• • asynchronous setcrtc, this property can be used to report any failures in that.
• */

+void drm_mode_connector_set_link_status_property(struct drm_connector *connector,

• 				 uint64_t link_status)
  

+{

• struct drm_device *dev = connector->dev;
• drm_modeset_lock(&dev->mode_config.connection_mutex, NULL);
• connector->state->link_status = link_status;
• drm_modeset_unlock(&dev->mode_config.connection_mutex);

+} +EXPORT_SYMBOL(drm_mode_connector_set_link_status_property);

int drm_mode_connector_set_obj_prop(struct drm_mode_object *obj, struct drm_property *property, uint64_t value) diff --git a/include/drm/drm_connector.h b/include/drm/drm_connector.h index 1218a0c..64e07c5 100644 --- a/include/drm/drm_connector.h +++ b/include/drm/drm_connector.h @@ -90,6 +90,17 @@ enum subpixel_order { };

/**

• • enum drm_link_status - connector's link_status property value
• • This enum is used as the connector's link status property value.
• • It is set to the values defined in uapi.
• */

+enum drm_link_status {

• DRM_LINK_STATUS_GOOD = DRM_MODE_LINK_STATUS_GOOD,
• DRM_LINK_STATUS_BAD = DRM_MODE_LINK_STATUS_BAD,

+};

+/**

• struct drm_display_info - runtime data about the connected sink
• Describes a given display (e.g. CRT or flat panel) and its limitations. For

@@ -213,6 +224,12 @@ struct drm_connector_state {

struct drm_encoder *best_encoder;

• /**

• * @link_status: Connector link_status to keep track of whether link is
  

• * GOOD or BAD to notify userspace if retraining is necessary.
  

• */
  

• enum drm_link_status link_status;
• struct drm_atomic_state *state;

};

@@ -776,6 +793,8 @@ int drm_mode_connector_set_path_property(struct drm_connector *connector, int drm_mode_connector_set_tile_property(struct drm_connector *connector); int drm_mode_connector_update_edid_property(struct drm_connector *connector, const struct edid *edid); +void drm_mode_connector_set_link_status_property(struct drm_connector *connector,

• 				 uint64_t link_status);
  
  

/**

• struct drm_tile_group - Tile group metadata

diff --git a/include/drm/drm_mode_config.h b/include/drm/drm_mode_config.h index bf9991b..86faee4 100644 --- a/include/drm/drm_mode_config.h +++ b/include/drm/drm_mode_config.h @@ -431,6 +431,11 @@ struct drm_mode_config { */ struct drm_property *tile_property; /**

• * @link_status_property: Default connector property for link status
  

• * of a connector
  

• */
  

• struct drm_property *link_status_property;
• /**
  • @plane_type_property: Default plane property to differentiate
  • CURSOR, PRIMARY and OVERLAY legacy uses of planes.
  */

diff --git a/include/uapi/drm/drm_mode.h b/include/uapi/drm/drm_mode.h index 728790b..309c478 100644 --- a/include/uapi/drm/drm_mode.h +++ b/include/uapi/drm/drm_mode.h @@ -123,6 +123,10 @@ #define DRM_MODE_DIRTY_ON 1 #define DRM_MODE_DIRTY_ANNOTATE 2

+/* Link Status options */ +#define DRM_MODE_LINK_STATUS_GOOD 0 +#define DRM_MODE_LINK_STATUS_BAD 1

struct drm_mode_modeinfo { __u32 clock; __u16 hdisplay; -- 1.9.1

---

Intel-gfx mailing list Intel-gfx@lists.freedesktop.org https://lists.freedesktop.org/mailman/listinfo/intel-gfx

At the time userspace does setcrtc, we've already promised the mode would work. The promise is based on the theoretical capabilities of the link, but it's possible we can't reach this in practice. The DP spec describes how the link should be reduced, but we can't reduce the link below the requirements of the mode. Black screen follows.

One idea would be to have setcrtc return a failure. However, it already should not fail as the atomic checks have passed. It would also conflict with the idea of making setcrtc asynchronous in the future, returning before the actual mode setting and link training.

Another idea is to train the link "upfront" at hotplug time, before pruning the mode list, so that we can do the pruning based on practical not theoretical capabilities. However, the changes for link training are pretty drastic, all for the sake of error handling and DP compliance, when the most common happy day scenario is the current approach of link training at mode setting time, using the optimal parameters for the mode. It is also not certain all hardware could do this without the pipe on; not even all our hardware can do this. Some of this can be solved, but not trivially.

Both of the above ideas also fail to address link degradation *during* operation.

The solution is to add a new "link-status" connector property in order to address link training failure in a way that: a) changes the current happy day scenario as little as possible, to avoid regressions, b) can be implemented the same way by all drm drivers, c) is still opt-in for the drivers and userspace, and opting out doesn't regress the user experience, d) doesn't prevent drivers from implementing better or alternate approaches, possibly without userspace involvement. And, of course, handles all the issues presented. In the usual happy day scenario, this is always "good". If something fails during or after a mode set, the kernel driver can set the link status to "bad" and issue a hotplug uevent for userspace to have it re-check the valid modes through GET_CONNECTOR IOCTL, and try modeset again. If the theoretical capabilities of the link can't be reached, the mode list is trimmed based on that.

v4: * Add comments in kernel-doc format (Daniel Vetter) * Update the kernel-doc for link-status (Sean Paul) v3: * Fixed a build error (Jani Saarinen) v2: * Removed connector->link_status (Daniel Vetter) * Set connector->state->link_status in drm_mode_connector_set_link_status_property (Daniel Vetter) * Set the connector_changed flag to true if connector->state->link_status changed. * Reset link_status to GOOD in update_output_state (Daniel Vetter) * Never allow userspace to set link status from Good To Bad (Daniel Vetter)

Reviewed-by: Sean Paul seanpaul@chromium.org Reviewed-by: Daniel Vetter daniel.vetter@ffwll.ch Acked-by: Tony Cheng tony.cheng@amd.com Acked-by: Harry Wentland harry.wentland@amd.com Cc: Jani Nikula jani.nikula@linux.intel.com Cc: Daniel Vetter daniel.vetter@intel.com Cc: Ville Syrjala ville.syrjala@linux.intel.com Cc: Chris Wilson chris@chris-wilson.co.uk Cc: Sean Paul seanpaul@chromium.org Signed-off-by: Manasi Navare manasi.d.navare@intel.com --- drivers/gpu/drm/drm_atomic.c | 10 +++++++ drivers/gpu/drm/drm_atomic_helper.c | 15 +++++++++++ drivers/gpu/drm/drm_connector.c | 53 +++++++++++++++++++++++++++++++++++++ include/drm/drm_connector.h | 19 +++++++++++++ include/drm/drm_mode_config.h | 5 ++++ include/uapi/drm/drm_mode.h | 4 +++ 6 files changed, 106 insertions(+)

diff --git a/drivers/gpu/drm/drm_atomic.c b/drivers/gpu/drm/drm_atomic.c index 19d7bcb..70a5152 100644 --- a/drivers/gpu/drm/drm_atomic.c +++ b/drivers/gpu/drm/drm_atomic.c @@ -1087,6 +1087,14 @@ int drm_atomic_connector_set_property(struct drm_connector *connector, * now?) atomic writes to DPMS property: */ return -EINVAL; + } else if (property == config->link_status_property) { + /* Never downgrade from GOOD to BAD on userspace's request here, + * only hw issues can do that. + */ + if (state->link_status == DRM_LINK_STATUS_GOOD) + return 0; + state->link_status = val; + return 0; } else if (connector->funcs->atomic_set_property) { return connector->funcs->atomic_set_property(connector, state, property, val); @@ -1135,6 +1143,8 @@ static void drm_atomic_connector_print_state(struct drm_printer *p, *val = (state->crtc) ? state->crtc->base.id : 0; } else if (property == config->dpms_property) { *val = connector->dpms; + } else if (property == config->link_status_property) { + *val = state->link_status; } else if (connector->funcs->atomic_get_property) { return connector->funcs->atomic_get_property(connector, state, property, val); diff --git a/drivers/gpu/drm/drm_atomic_helper.c b/drivers/gpu/drm/drm_atomic_helper.c index 494680c..3d6c1ce 100644 --- a/drivers/gpu/drm/drm_atomic_helper.c +++ b/drivers/gpu/drm/drm_atomic_helper.c @@ -519,6 +519,13 @@ static int handle_conflicting_encoders(struct drm_atomic_state *state, connector_state); if (ret) return ret; + if (connector->state->crtc) { + crtc_state = drm_atomic_get_existing_crtc_state(state, + connector->state->crtc); + if (connector->state->link_status != + connector_state->link_status) + crtc_state->connectors_changed = true; + } }

/* @@ -2258,6 +2265,8 @@ static int update_output_state(struct drm_atomic_state *state, NULL); if (ret) return ret; + /* Make sure legacy setCrtc always re-trains */ + conn_state->link_status = DRM_LINK_STATUS_GOOD; } }

@@ -2301,6 +2310,12 @@ static int update_output_state(struct drm_atomic_state *state, * * Provides a default crtc set_config handler using the atomic driver interface. * + * NOTE: For backwards compatibility with old userspace this automatically + * resets the "link-status" property to GOOD, to force any link + * re-training. The SETCRTC ioctl does not define whether an update does + * need a full modeset or just a plane update, hence we're allowed to do + * that. See also drm_mode_connector_set_link_status_property(). + * * Returns: * Returns 0 on success, negative errno numbers on failure. */ diff --git a/drivers/gpu/drm/drm_connector.c b/drivers/gpu/drm/drm_connector.c index 5a45262..eee038a 100644 --- a/drivers/gpu/drm/drm_connector.c +++ b/drivers/gpu/drm/drm_connector.c @@ -243,6 +243,10 @@ int drm_connector_init(struct drm_device *dev, drm_object_attach_property(&connector->base, config->dpms_property, 0);

+ drm_object_attach_property(&connector->base, + config->link_status_property, + 0); + if (drm_core_check_feature(dev, DRIVER_ATOMIC)) { drm_object_attach_property(&connector->base, config->prop_crtc_id, 0); } @@ -506,6 +510,12 @@ const char *drm_get_subpixel_order_name(enum subpixel_order order) }; DRM_ENUM_NAME_FN(drm_get_dpms_name, drm_dpms_enum_list)

+static const struct drm_prop_enum_list drm_link_status_enum_list[] = { + { DRM_MODE_LINK_STATUS_GOOD, "Good" }, + { DRM_MODE_LINK_STATUS_BAD, "Bad" }, +}; +DRM_ENUM_NAME_FN(drm_get_link_status_name, drm_link_status_enum_list) + /** * drm_display_info_set_bus_formats - set the supported bus formats * @info: display info to store bus formats in @@ -625,6 +635,11 @@ int drm_display_info_set_bus_formats(struct drm_display_info *info, * tiling and virtualize both &drm_crtc and &drm_plane if needed. Drivers * should update this value using drm_mode_connector_set_tile_property(). * Userspace cannot change this property. + * link-status: + * Connector link-status property to indicate the status of link. The default + * value of link-status is "GOOD". If something fails during or after modeset, + * the kernel driver may set this to "BAD" and issue a hotplug uevent. Drivers + * should update this value using drm_mode_connector_set_link_status_property(). * * Connectors also have one standardized atomic property: * @@ -666,6 +681,13 @@ int drm_connector_create_standard_properties(struct drm_device *dev) return -ENOMEM; dev->mode_config.tile_property = prop;

+ prop = drm_property_create_enum(dev, 0, "link-status", + drm_link_status_enum_list, + ARRAY_SIZE(drm_link_status_enum_list)); + if (!prop) + return -ENOMEM; + dev->mode_config.link_status_property = prop; + return 0; }

@@ -995,6 +1017,37 @@ int drm_mode_connector_update_edid_property(struct drm_connector *connector, } EXPORT_SYMBOL(drm_mode_connector_update_edid_property);

+/** + * drm_mode_connector_set_link_status_property - Set link status property of a connector + * @connector: drm connector + * @link_status: new value of link status property (0: Good, 1: Bad) + * + * In usual working scenario, this link status property will always be set to + * "GOOD". If something fails during or after a mode set, the kernel driver + * may set this link status property to "BAD". The caller then needs to send a + * hotplug uevent for userspace to re-check the valid modes through + * GET_CONNECTOR_IOCTL and retry modeset. + * + * Note: Drivers cannot rely on userspace to support this property and + * issue a modeset. As such, they may choose to handle issues (like + * re-training a link) without userspace's intervention. + * + * The reason for adding this property is to handle link training failures, but + * it is not limited to DP or link training. For example, if we implement + * asynchronous setcrtc, this property can be used to report any failures in that. + */ +void drm_mode_connector_set_link_status_property(struct drm_connector *connector, + uint64_t link_status) +{ + struct drm_device *dev = connector->dev; + + drm_modeset_lock(&dev->mode_config.connection_mutex, NULL); + connector->state->link_status = link_status; + drm_modeset_unlock(&dev->mode_config.connection_mutex); + +} +EXPORT_SYMBOL(drm_mode_connector_set_link_status_property); + int drm_mode_connector_set_obj_prop(struct drm_mode_object *obj, struct drm_property *property, uint64_t value) diff --git a/include/drm/drm_connector.h b/include/drm/drm_connector.h index 1218a0c..64e07c5 100644 --- a/include/drm/drm_connector.h +++ b/include/drm/drm_connector.h @@ -90,6 +90,17 @@ enum subpixel_order { };

/** + * enum drm_link_status - connector's link_status property value + * + * This enum is used as the connector's link status property value. + * It is set to the values defined in uapi. + */ +enum drm_link_status { + DRM_LINK_STATUS_GOOD = DRM_MODE_LINK_STATUS_GOOD, + DRM_LINK_STATUS_BAD = DRM_MODE_LINK_STATUS_BAD, +}; + +/** * struct drm_display_info - runtime data about the connected sink * * Describes a given display (e.g. CRT or flat panel) and its limitations. For @@ -213,6 +224,12 @@ struct drm_connector_state {

struct drm_encoder *best_encoder;

+ /** + * @link_status: Connector link_status to keep track of whether link is + * GOOD or BAD to notify userspace if retraining is necessary. + */ + enum drm_link_status link_status; + struct drm_atomic_state *state; };

@@ -776,6 +793,8 @@ int drm_mode_connector_set_path_property(struct drm_connector *connector, int drm_mode_connector_set_tile_property(struct drm_connector *connector); int drm_mode_connector_update_edid_property(struct drm_connector *connector, const struct edid *edid); +void drm_mode_connector_set_link_status_property(struct drm_connector *connector, + uint64_t link_status);

/** * struct drm_tile_group - Tile group metadata diff --git a/include/drm/drm_mode_config.h b/include/drm/drm_mode_config.h index bf9991b..86faee4 100644 --- a/include/drm/drm_mode_config.h +++ b/include/drm/drm_mode_config.h @@ -431,6 +431,11 @@ struct drm_mode_config { */ struct drm_property *tile_property; /** + * @link_status_property: Default connector property for link status + * of a connector + */ + struct drm_property *link_status_property; + /** * @plane_type_property: Default plane property to differentiate * CURSOR, PRIMARY and OVERLAY legacy uses of planes. */ diff --git a/include/uapi/drm/drm_mode.h b/include/uapi/drm/drm_mode.h index 728790b..309c478 100644 --- a/include/uapi/drm/drm_mode.h +++ b/include/uapi/drm/drm_mode.h @@ -123,6 +123,10 @@ #define DRM_MODE_DIRTY_ON 1 #define DRM_MODE_DIRTY_ANNOTATE 2

+/* Link Status options */ +#define DRM_MODE_LINK_STATUS_GOOD 0 +#define DRM_MODE_LINK_STATUS_BAD 1 + struct drm_mode_modeinfo { __u32 clock; __u16 hdisplay;

On Thu, Dec 08, 2016 at 05:05:54PM +0200, Jani Nikula wrote:

On Tue, 06 Dec 2016, Manasi Navare manasi.d.navare@intel.com wrote:

...

At the time userspace does setcrtc, we've already promised the mode would work. The promise is based on the theoretical capabilities of the link, but it's possible we can't reach this in practice. The DP spec describes how the link should be reduced, but we can't reduce the link below the requirements of the mode. Black screen follows.

One idea would be to have setcrtc return a failure. However, it already should not fail as the atomic checks have passed. It would also conflict with the idea of making setcrtc asynchronous in the future, returning before the actual mode setting and link training.

Another idea is to train the link "upfront" at hotplug time, before pruning the mode list, so that we can do the pruning based on practical not theoretical capabilities. However, the changes for link training are pretty drastic, all for the sake of error handling and DP compliance, when the most common happy day scenario is the current approach of link training at mode setting time, using the optimal parameters for the mode. It is also not certain all hardware could do this without the pipe on; not even all our hardware can do this. Some of this can be solved, but not trivially.

Both of the above ideas also fail to address link degradation *during* operation.

The solution is to add a new "link-status" connector property in order to address link training failure in a way that: a) changes the current happy day scenario as little as possible, to avoid regressions, b) can be implemented the same way by all drm drivers, c) is still opt-in for the drivers and userspace, and opting out doesn't regress the user experience, d) doesn't prevent drivers from implementing better or alternate approaches, possibly without userspace involvement. And, of course, handles all the issues presented. In the usual happy day scenario, this is always "good". If something fails during or after a mode set, the kernel driver can set the link status to "bad" and issue a hotplug uevent for userspace to have it re-check the valid modes through GET_CONNECTOR IOCTL, and try modeset again. If the theoretical capabilities of the link can't be reached, the mode list is trimmed based on that.

v4:

• Add comments in kernel-doc format (Daniel Vetter)
• Update the kernel-doc for link-status (Sean Paul)

v3:

• Fixed a build error (Jani Saarinen)

v2:

• Removed connector->link_status (Daniel Vetter)
• Set connector->state->link_status in drm_mode_connector_set_link_status_property

(Daniel Vetter)

• Set the connector_changed flag to true if connector->state->link_status changed.
• Reset link_status to GOOD in update_output_state (Daniel Vetter)
• Never allow userspace to set link status from Good To Bad (Daniel Vetter)

Reviewed-by: Sean Paul seanpaul@chromium.org Reviewed-by: Daniel Vetter daniel.vetter@ffwll.ch Acked-by: Tony Cheng tony.cheng@amd.com Acked-by: Harry Wentland harry.wentland@amd.com Cc: Jani Nikula jani.nikula@linux.intel.com Cc: Daniel Vetter daniel.vetter@intel.com Cc: Ville Syrjala ville.syrjala@linux.intel.com Cc: Chris Wilson chris@chris-wilson.co.uk Cc: Sean Paul seanpaul@chromium.org Signed-off-by: Manasi Navare manasi.d.navare@intel.com

---

drivers/gpu/drm/drm_atomic.c | 10 +++++++ drivers/gpu/drm/drm_atomic_helper.c | 15 +++++++++++ drivers/gpu/drm/drm_connector.c | 53 +++++++++++++++++++++++++++++++++++++ include/drm/drm_connector.h | 19 +++++++++++++ include/drm/drm_mode_config.h | 5 ++++ include/uapi/drm/drm_mode.h | 4 +++ 6 files changed, 106 insertions(+)

diff --git a/drivers/gpu/drm/drm_atomic.c b/drivers/gpu/drm/drm_atomic.c index 19d7bcb..70a5152 100644 --- a/drivers/gpu/drm/drm_atomic.c +++ b/drivers/gpu/drm/drm_atomic.c @@ -1087,6 +1087,14 @@ int drm_atomic_connector_set_property(struct drm_connector *connector, * now?) atomic writes to DPMS property: */ return -EINVAL;

• } else if (property == config->link_status_property) {

• /* Never downgrade from GOOD to BAD on userspace's request here,
  

•  * only hw issues can do that.
  

•  */
  

• if (state->link_status == DRM_LINK_STATUS_GOOD)
  

• 	return 0;
  

Is it conceivable we'd want to support the userspace requesting the change from GOOD to BAD at a later point in time? If yes, perhaps the right answer here is to return an error on GOOD->BAD transition, so the userspace can distinguish between the kernel rejecting vs. accepting the property set?

One of the design ideas behind atomic is that userspace doesn't need to be able to understand all properties, but still needs to be able to restore the state it wants on VT switch. Or as a "pls reset my display to boot-up state because my compositor made a mess of it". That's why e.g. the various fence ptr properties read back as 0, and the fence fd input parameters read back as -1, to make sure dumb save/restore doesn't do something stupid.

Same idea here for silently rejecting a good->bad transition: We don't want userspace to accidentally break the display when it restores state. At least that's been my thinking behind suggesting this. Might want to document that somewhere a bit better.

BR, Jani.

...

• state->link_status = val;
  

• return 0;
  

  } else if (connector->funcs->atomic_set_property) { return connector->funcs->atomic_set_property(connector, state, property, val);

@@ -1135,6 +1143,8 @@ static void drm_atomic_connector_print_state(struct drm_printer *p, *val = (state->crtc) ? state->crtc->base.id : 0; } else if (property == config->dpms_property) { *val = connector->dpms;

• } else if (property == config->link_status_property) {

• *val = state->link_status;
  

  } else if (connector->funcs->atomic_get_property) { return connector->funcs->atomic_get_property(connector, state, property, val);

diff --git a/drivers/gpu/drm/drm_atomic_helper.c b/drivers/gpu/drm/drm_atomic_helper.c index 494680c..3d6c1ce 100644 --- a/drivers/gpu/drm/drm_atomic_helper.c +++ b/drivers/gpu/drm/drm_atomic_helper.c @@ -519,6 +519,13 @@ static int handle_conflicting_encoders(struct drm_atomic_state *state, connector_state); if (ret) return ret;

• if (connector->state->crtc) {
  

• 	crtc_state = drm_atomic_get_existing_crtc_state(state,
  

• 							connector->state->crtc);
  

• 	if (connector->state->link_status !=
  

• 	    connector_state->link_status)
  

• 		crtc_state->connectors_changed = true;
  

• }
  

  }

  /*

@@ -2258,6 +2265,8 @@ static int update_output_state(struct drm_atomic_state *state, NULL); if (ret) return ret;

• 	/* Make sure legacy setCrtc always re-trains */
  

• 	conn_state->link_status = DRM_LINK_STATUS_GOOD;
  

  } }

@@ -2301,6 +2310,12 @@ static int update_output_state(struct drm_atomic_state *state,

• Provides a default crtc set_config handler using the atomic driver interface.

• • NOTE: For backwards compatibility with old userspace this automatically
• • resets the "link-status" property to GOOD, to force any link
• • re-training. The SETCRTC ioctl does not define whether an update does
• • need a full modeset or just a plane update, hence we're allowed to do
• • that. See also drm_mode_connector_set_link_status_property().
• • Returns:
  • Returns 0 on success, negative errno numbers on failure.
  */

diff --git a/drivers/gpu/drm/drm_connector.c b/drivers/gpu/drm/drm_connector.c index 5a45262..eee038a 100644 --- a/drivers/gpu/drm/drm_connector.c +++ b/drivers/gpu/drm/drm_connector.c @@ -243,6 +243,10 @@ int drm_connector_init(struct drm_device *dev, drm_object_attach_property(&connector->base, config->dpms_property, 0);

• drm_object_attach_property(&connector->base,

• 		   config->link_status_property,
  

• 		   0);
  

• if (drm_core_check_feature(dev, DRIVER_ATOMIC)) { drm_object_attach_property(&connector->base, config->prop_crtc_id, 0); }

@@ -506,6 +510,12 @@ const char *drm_get_subpixel_order_name(enum subpixel_order order) }; DRM_ENUM_NAME_FN(drm_get_dpms_name, drm_dpms_enum_list)

+static const struct drm_prop_enum_list drm_link_status_enum_list[] = {

• { DRM_MODE_LINK_STATUS_GOOD, "Good" },
• { DRM_MODE_LINK_STATUS_BAD, "Bad" },

+}; +DRM_ENUM_NAME_FN(drm_get_link_status_name, drm_link_status_enum_list)

/**

• drm_display_info_set_bus_formats - set the supported bus formats
• @info: display info to store bus formats in

@@ -625,6 +635,11 @@ int drm_display_info_set_bus_formats(struct drm_display_info *info,

• tiling and virtualize both &drm_crtc and &drm_plane if needed. Drivers
• should update this value using drm_mode_connector_set_tile_property().
• Userspace cannot change this property.

• • link-status:

• •  Connector link-status property to indicate the status of link. The default
    

• •  value of link-status is "GOOD". If something fails during or after modeset,
    

• •  the kernel driver may set this to "BAD" and issue a hotplug uevent. Drivers
    

• •  should update this value using drm_mode_connector_set_link_status_property().
    

E.g. by adding here "Note that userspace trying to change from GOOD to BAD is silently rejected, to avoid a dumb save/restore for e.g. VT switching from upsetting the display state."

Does that sound good?

Cheers, Daniel

...

• Connectors also have one standardized atomic property:

@@ -666,6 +681,13 @@ int drm_connector_create_standard_properties(struct drm_device *dev) return -ENOMEM; dev->mode_config.tile_property = prop;

• prop = drm_property_create_enum(dev, 0, "link-status",

• 			drm_link_status_enum_list,
  

• 			ARRAY_SIZE(drm_link_status_enum_list));
  

• if (!prop)

• return -ENOMEM;
  

• dev->mode_config.link_status_property = prop;
• return 0;

}

@@ -995,6 +1017,37 @@ int drm_mode_connector_update_edid_property(struct drm_connector *connector, } EXPORT_SYMBOL(drm_mode_connector_update_edid_property);

+/**

• • drm_mode_connector_set_link_status_property - Set link status property of a connector
• • @connector: drm connector
• • @link_status: new value of link status property (0: Good, 1: Bad)
• • In usual working scenario, this link status property will always be set to
• • "GOOD". If something fails during or after a mode set, the kernel driver
• • may set this link status property to "BAD". The caller then needs to send a
• • hotplug uevent for userspace to re-check the valid modes through
• • GET_CONNECTOR_IOCTL and retry modeset.
• • Note: Drivers cannot rely on userspace to support this property and
• • issue a modeset. As such, they may choose to handle issues (like
• • re-training a link) without userspace's intervention.
• • The reason for adding this property is to handle link training failures, but
• • it is not limited to DP or link training. For example, if we implement
• • asynchronous setcrtc, this property can be used to report any failures in that.
• */

+void drm_mode_connector_set_link_status_property(struct drm_connector *connector,

• 				 uint64_t link_status)
  

+{

• struct drm_device *dev = connector->dev;
• drm_modeset_lock(&dev->mode_config.connection_mutex, NULL);
• connector->state->link_status = link_status;
• drm_modeset_unlock(&dev->mode_config.connection_mutex);

+} +EXPORT_SYMBOL(drm_mode_connector_set_link_status_property);

int drm_mode_connector_set_obj_prop(struct drm_mode_object *obj, struct drm_property *property, uint64_t value) diff --git a/include/drm/drm_connector.h b/include/drm/drm_connector.h index 1218a0c..64e07c5 100644 --- a/include/drm/drm_connector.h +++ b/include/drm/drm_connector.h @@ -90,6 +90,17 @@ enum subpixel_order { };

/**

• • enum drm_link_status - connector's link_status property value
• • This enum is used as the connector's link status property value.
• • It is set to the values defined in uapi.
• */

+enum drm_link_status {

• DRM_LINK_STATUS_GOOD = DRM_MODE_LINK_STATUS_GOOD,
• DRM_LINK_STATUS_BAD = DRM_MODE_LINK_STATUS_BAD,

+};

+/**

• struct drm_display_info - runtime data about the connected sink
• Describes a given display (e.g. CRT or flat panel) and its limitations. For

@@ -213,6 +224,12 @@ struct drm_connector_state {

struct drm_encoder *best_encoder;

• /**

• * @link_status: Connector link_status to keep track of whether link is
  

• * GOOD or BAD to notify userspace if retraining is necessary.
  

• */
  

• enum drm_link_status link_status;
• struct drm_atomic_state *state;

};

@@ -776,6 +793,8 @@ int drm_mode_connector_set_path_property(struct drm_connector *connector, int drm_mode_connector_set_tile_property(struct drm_connector *connector); int drm_mode_connector_update_edid_property(struct drm_connector *connector, const struct edid *edid); +void drm_mode_connector_set_link_status_property(struct drm_connector *connector,

• 				 uint64_t link_status);
  
  

/**

• struct drm_tile_group - Tile group metadata

diff --git a/include/drm/drm_mode_config.h b/include/drm/drm_mode_config.h index bf9991b..86faee4 100644 --- a/include/drm/drm_mode_config.h +++ b/include/drm/drm_mode_config.h @@ -431,6 +431,11 @@ struct drm_mode_config { */ struct drm_property *tile_property; /**

• * @link_status_property: Default connector property for link status
  

• * of a connector
  

• */
  

• struct drm_property *link_status_property;
• /**
  • @plane_type_property: Default plane property to differentiate
  • CURSOR, PRIMARY and OVERLAY legacy uses of planes.
  */

diff --git a/include/uapi/drm/drm_mode.h b/include/uapi/drm/drm_mode.h index 728790b..309c478 100644 --- a/include/uapi/drm/drm_mode.h +++ b/include/uapi/drm/drm_mode.h @@ -123,6 +123,10 @@ #define DRM_MODE_DIRTY_ON 1 #define DRM_MODE_DIRTY_ANNOTATE 2

+/* Link Status options */ +#define DRM_MODE_LINK_STATUS_GOOD 0 +#define DRM_MODE_LINK_STATUS_BAD 1

struct drm_mode_modeinfo { __u32 clock; __u16 hdisplay;

-- Jani Nikula, Intel Open Source Technology Center _______________________________________________ Intel-gfx mailing list Intel-gfx@lists.freedesktop.org https://lists.freedesktop.org/mailman/listinfo/intel-gfx

<snip>

diff --git a/drivers/gpu/drm/drm_atomic.c b/drivers/gpu/drm/drm_atomic.c index 72fa5b2..34eddb9 100644 --- a/drivers/gpu/drm/drm_atomic.c +++ b/drivers/gpu/drm/drm_atomic.c @@ -1087,6 +1087,22 @@ int drm_atomic_connector_set_property(struct drm_connector *connector, * now?) atomic writes to DPMS property: */ return -EINVAL;

•   } else if (property == config->link_status_property) {
  

•           /* Never downgrade from GOOD to BAD on userspace's request here,
  

•            * only hw issues can do that.
  

•            *
  

•            * For an atomic property the userspace doesn't need to be able
  

•            * to understand all the properties, but needs to be able to
  

•            * restore the state itw ants on VT switch. So if the userspace
  

s/itw ants/it wants/

<snip>

At the time userspace does setcrtc, we've already promised the mode would work. The promise is based on the theoretical capabilities of the link, but it's possible we can't reach this in practice. The DP spec describes how the link should be reduced, but we can't reduce the link below the requirements of the mode. Black screen follows.

One idea would be to have setcrtc return a failure. However, it already should not fail as the atomic checks have passed. It would also conflict with the idea of making setcrtc asynchronous in the future, returning before the actual mode setting and link training.

Another idea is to train the link "upfront" at hotplug time, before pruning the mode list, so that we can do the pruning based on practical not theoretical capabilities. However, the changes for link training are pretty drastic, all for the sake of error handling and DP compliance, when the most common happy day scenario is the current approach of link training at mode setting time, using the optimal parameters for the mode. It is also not certain all hardware could do this without the pipe on; not even all our hardware can do this. Some of this can be solved, but not trivially.

Both of the above ideas also fail to address link degradation *during* operation.

The solution is to add a new "link-status" connector property in order to address link training failure in a way that: a) changes the current happy day scenario as little as possible, to avoid regressions, b) can be implemented the same way by all drm drivers, c) is still opt-in for the drivers and userspace, and opting out doesn't regress the user experience, d) doesn't prevent drivers from implementing better or alternate approaches, possibly without userspace involvement. And, of course, handles all the issues presented. In the usual happy day scenario, this is always "good". If something fails during or after a mode set, the kernel driver can set the link status to "bad" and issue a hotplug uevent for userspace to have it re-check the valid modes through GET_CONNECTOR IOCTL, and try modeset again. If the theoretical capabilities of the link can't be reached, the mode list is trimmed based on that.

v6: * Fix a typo in kernel doc (Sean Paul) v5: * Clarify doc for silent rejection of atomic properties by driver (Daniel Vetter) v4: * Add comments in kernel-doc format (Daniel Vetter) * Update the kernel-doc for link-status (Sean Paul) v3: * Fixed a build error (Jani Saarinen) v2: * Removed connector->link_status (Daniel Vetter) * Set connector->state->link_status in drm_mode_connector_set_link_status_property (Daniel Vetter) * Set the connector_changed flag to true if connector->state->link_status changed. * Reset link_status to GOOD in update_output_state (Daniel Vetter) * Never allow userspace to set link status from Good To Bad (Daniel Vetter)

Reviewed-by: Sean Paul seanpaul@chromium.org Reviewed-by: Daniel Vetter daniel.vetter@ffwll.ch Acked-by: Tony Cheng tony.cheng@amd.com Acked-by: Harry Wentland harry.wentland@amd.com Cc: Jani Nikula jani.nikula@linux.intel.com Cc: Daniel Vetter daniel.vetter@intel.com Cc: Ville Syrjala ville.syrjala@linux.intel.com Cc: Chris Wilson chris@chris-wilson.co.uk Cc: Sean Paul seanpaul@chromium.org Signed-off-by: Manasi Navare manasi.d.navare@intel.com --- drivers/gpu/drm/drm_atomic.c | 18 +++++++++++++ drivers/gpu/drm/drm_atomic_helper.c | 15 +++++++++++ drivers/gpu/drm/drm_connector.c | 53 +++++++++++++++++++++++++++++++++++++ include/drm/drm_connector.h | 19 +++++++++++++ include/drm/drm_mode_config.h | 5 ++++ include/uapi/drm/drm_mode.h | 4 +++ 6 files changed, 114 insertions(+)

diff --git a/drivers/gpu/drm/drm_atomic.c b/drivers/gpu/drm/drm_atomic.c index 72fa5b2..673f3af 100644 --- a/drivers/gpu/drm/drm_atomic.c +++ b/drivers/gpu/drm/drm_atomic.c @@ -1087,6 +1087,22 @@ int drm_atomic_connector_set_property(struct drm_connector *connector, * now?) atomic writes to DPMS property: */ return -EINVAL; + } else if (property == config->link_status_property) { + /* Never downgrade from GOOD to BAD on userspace's request here, + * only hw issues can do that. + * + * For an atomic property the userspace doesn't need to be able + * to understand all the properties, but needs to be able to + * restore the state it wants on VT switch. So if the userspace + * tries to change the link_status from GOOD to BAD, driver + * silently rejects it and returns a 0. This prevents userspace + * from accidently breaking the display when it restores the + * state. + */ + if (state->link_status == DRM_LINK_STATUS_GOOD) + return 0; + state->link_status = val; + return 0; } else if (connector->funcs->atomic_set_property) { return connector->funcs->atomic_set_property(connector, state, property, val); @@ -1135,6 +1151,8 @@ static void drm_atomic_connector_print_state(struct drm_printer *p, *val = (state->crtc) ? state->crtc->base.id : 0; } else if (property == config->dpms_property) { *val = connector->dpms; + } else if (property == config->link_status_property) { + *val = state->link_status; } else if (connector->funcs->atomic_get_property) { return connector->funcs->atomic_get_property(connector, state, property, val); diff --git a/drivers/gpu/drm/drm_atomic_helper.c b/drivers/gpu/drm/drm_atomic_helper.c index 23767df..5e7af63 100644 --- a/drivers/gpu/drm/drm_atomic_helper.c +++ b/drivers/gpu/drm/drm_atomic_helper.c @@ -519,6 +519,13 @@ static int handle_conflicting_encoders(struct drm_atomic_state *state, connector_state); if (ret) return ret; + if (connector->state->crtc) { + crtc_state = drm_atomic_get_existing_crtc_state(state, + connector->state->crtc); + if (connector->state->link_status != + connector_state->link_status) + crtc_state->connectors_changed = true; + } }

/* @@ -2255,6 +2262,8 @@ static int update_output_state(struct drm_atomic_state *state, NULL); if (ret) return ret; + /* Make sure legacy setCrtc always re-trains */ + conn_state->link_status = DRM_LINK_STATUS_GOOD; } }

@@ -2298,6 +2307,12 @@ static int update_output_state(struct drm_atomic_state *state, * * Provides a default crtc set_config handler using the atomic driver interface. * + * NOTE: For backwards compatibility with old userspace this automatically + * resets the "link-status" property to GOOD, to force any link + * re-training. The SETCRTC ioctl does not define whether an update does + * need a full modeset or just a plane update, hence we're allowed to do + * that. See also drm_mode_connector_set_link_status_property(). + * * Returns: * Returns 0 on success, negative errno numbers on failure. */ diff --git a/drivers/gpu/drm/drm_connector.c b/drivers/gpu/drm/drm_connector.c index 5a45262..eee038a 100644 --- a/drivers/gpu/drm/drm_connector.c +++ b/drivers/gpu/drm/drm_connector.c @@ -243,6 +243,10 @@ int drm_connector_init(struct drm_device *dev, drm_object_attach_property(&connector->base, config->dpms_property, 0);

+ drm_object_attach_property(&connector->base, + config->link_status_property, + 0); + if (drm_core_check_feature(dev, DRIVER_ATOMIC)) { drm_object_attach_property(&connector->base, config->prop_crtc_id, 0); } @@ -506,6 +510,12 @@ const char *drm_get_subpixel_order_name(enum subpixel_order order) }; DRM_ENUM_NAME_FN(drm_get_dpms_name, drm_dpms_enum_list)

+static const struct drm_prop_enum_list drm_link_status_enum_list[] = { + { DRM_MODE_LINK_STATUS_GOOD, "Good" }, + { DRM_MODE_LINK_STATUS_BAD, "Bad" }, +}; +DRM_ENUM_NAME_FN(drm_get_link_status_name, drm_link_status_enum_list) + /** * drm_display_info_set_bus_formats - set the supported bus formats * @info: display info to store bus formats in @@ -625,6 +635,11 @@ int drm_display_info_set_bus_formats(struct drm_display_info *info, * tiling and virtualize both &drm_crtc and &drm_plane if needed. Drivers * should update this value using drm_mode_connector_set_tile_property(). * Userspace cannot change this property. + * link-status: + * Connector link-status property to indicate the status of link. The default + * value of link-status is "GOOD". If something fails during or after modeset, + * the kernel driver may set this to "BAD" and issue a hotplug uevent. Drivers + * should update this value using drm_mode_connector_set_link_status_property(). * * Connectors also have one standardized atomic property: * @@ -666,6 +681,13 @@ int drm_connector_create_standard_properties(struct drm_device *dev) return -ENOMEM; dev->mode_config.tile_property = prop;

+ prop = drm_property_create_enum(dev, 0, "link-status", + drm_link_status_enum_list, + ARRAY_SIZE(drm_link_status_enum_list)); + if (!prop) + return -ENOMEM; + dev->mode_config.link_status_property = prop; + return 0; }

@@ -995,6 +1017,37 @@ int drm_mode_connector_update_edid_property(struct drm_connector *connector, } EXPORT_SYMBOL(drm_mode_connector_update_edid_property);

+/** + * drm_mode_connector_set_link_status_property - Set link status property of a connector + * @connector: drm connector + * @link_status: new value of link status property (0: Good, 1: Bad) + * + * In usual working scenario, this link status property will always be set to + * "GOOD". If something fails during or after a mode set, the kernel driver + * may set this link status property to "BAD". The caller then needs to send a + * hotplug uevent for userspace to re-check the valid modes through + * GET_CONNECTOR_IOCTL and retry modeset. + * + * Note: Drivers cannot rely on userspace to support this property and + * issue a modeset. As such, they may choose to handle issues (like + * re-training a link) without userspace's intervention. + * + * The reason for adding this property is to handle link training failures, but + * it is not limited to DP or link training. For example, if we implement + * asynchronous setcrtc, this property can be used to report any failures in that. + */ +void drm_mode_connector_set_link_status_property(struct drm_connector *connector, + uint64_t link_status) +{ + struct drm_device *dev = connector->dev; + + drm_modeset_lock(&dev->mode_config.connection_mutex, NULL); + connector->state->link_status = link_status; + drm_modeset_unlock(&dev->mode_config.connection_mutex); + +} +EXPORT_SYMBOL(drm_mode_connector_set_link_status_property); + int drm_mode_connector_set_obj_prop(struct drm_mode_object *obj, struct drm_property *property, uint64_t value) diff --git a/include/drm/drm_connector.h b/include/drm/drm_connector.h index 1218a0c..64e07c5 100644 --- a/include/drm/drm_connector.h +++ b/include/drm/drm_connector.h @@ -90,6 +90,17 @@ enum subpixel_order { };

/** + * enum drm_link_status - connector's link_status property value + * + * This enum is used as the connector's link status property value. + * It is set to the values defined in uapi. + */ +enum drm_link_status { + DRM_LINK_STATUS_GOOD = DRM_MODE_LINK_STATUS_GOOD, + DRM_LINK_STATUS_BAD = DRM_MODE_LINK_STATUS_BAD, +}; + +/** * struct drm_display_info - runtime data about the connected sink * * Describes a given display (e.g. CRT or flat panel) and its limitations. For @@ -213,6 +224,12 @@ struct drm_connector_state {

struct drm_encoder *best_encoder;

+ /** + * @link_status: Connector link_status to keep track of whether link is + * GOOD or BAD to notify userspace if retraining is necessary. + */ + enum drm_link_status link_status; + struct drm_atomic_state *state; };

@@ -776,6 +793,8 @@ int drm_mode_connector_set_path_property(struct drm_connector *connector, int drm_mode_connector_set_tile_property(struct drm_connector *connector); int drm_mode_connector_update_edid_property(struct drm_connector *connector, const struct edid *edid); +void drm_mode_connector_set_link_status_property(struct drm_connector *connector, + uint64_t link_status);

/** * struct drm_tile_group - Tile group metadata diff --git a/include/drm/drm_mode_config.h b/include/drm/drm_mode_config.h index bf9991b..86faee4 100644 --- a/include/drm/drm_mode_config.h +++ b/include/drm/drm_mode_config.h @@ -431,6 +431,11 @@ struct drm_mode_config { */ struct drm_property *tile_property; /** + * @link_status_property: Default connector property for link status + * of a connector + */ + struct drm_property *link_status_property; + /** * @plane_type_property: Default plane property to differentiate * CURSOR, PRIMARY and OVERLAY legacy uses of planes. */ diff --git a/include/uapi/drm/drm_mode.h b/include/uapi/drm/drm_mode.h index 728790b..309c478 100644 --- a/include/uapi/drm/drm_mode.h +++ b/include/uapi/drm/drm_mode.h @@ -123,6 +123,10 @@ #define DRM_MODE_DIRTY_ON 1 #define DRM_MODE_DIRTY_ANNOTATE 2

+/* Link Status options */ +#define DRM_MODE_LINK_STATUS_GOOD 0 +#define DRM_MODE_LINK_STATUS_BAD 1 + struct drm_mode_modeinfo { __u32 clock; __u16 hdisplay;

Jani,

Could you please review this patch? This is the patch that calculates the max sink link rate and max sink lane count only once at hotplug and then anytime the max lane count and common rates are requested, the helper functions use these values. This simplifies the fallback logic since we can go ahead and update the max sink link rate and lane count to these fallback lower values since link training failed so thus capping the max values to fallback values. This is better than defining common_rates array in intel_dp, because its easier to update the max link rate vs going and updating and resizing the whole array. Also with this approach the helper functions do not change.

Regards Manasi

On Mon, Dec 05, 2016 at 04:27:36PM -0800, Manasi Navare wrote:

Sink's capabilities are advertised through DPCD registers and get updated only on hotplug. So they should be computed only once in the long pulse handler and saved off in intel_dp structure for the use later. For this reason two new fields max_sink_lane_count and max_sink_link_bw are added to intel_dp structure.

This also simplifies the fallback link rate/lane count logic to handle link training failure. In that case, the max_sink_link_bw and max_sink_lane_count can be reccomputed to match the fallback values lowering the sink capabilities due to link train failure.

Cc: Ville Syrjala ville.syrjala@linux.intel.com Cc: Jani Nikula jani.nikula@linux.intel.com Cc: Daniel Vetter daniel.vetter@intel.com Signed-off-by: Manasi Navare manasi.d.navare@intel.com

---

drivers/gpu/drm/i915/intel_dp.c | 10 ++++++++-- drivers/gpu/drm/i915/intel_drv.h | 4 ++++ 2 files changed, 12 insertions(+), 2 deletions(-)

diff --git a/drivers/gpu/drm/i915/intel_dp.c b/drivers/gpu/drm/i915/intel_dp.c index db75bb9..434dc7d 100644 --- a/drivers/gpu/drm/i915/intel_dp.c +++ b/drivers/gpu/drm/i915/intel_dp.c @@ -156,7 +156,7 @@ static u8 intel_dp_max_lane_count(struct intel_dp *intel_dp) u8 source_max, sink_max;

source_max = intel_dig_port->max_lanes;

• sink_max = drm_dp_max_lane_count(intel_dp->dpcd);

• sink_max = intel_dp->max_sink_lane_count;

  return min(source_max, sink_max);

} @@ -213,7 +213,7 @@ static u8 intel_dp_max_lane_count(struct intel_dp *intel_dp)

*sink_rates = default_rates;

• return (intel_dp_max_link_bw(intel_dp) >> 3) + 1;

• return (intel_dp->max_sink_link_bw >> 3) + 1;

}

static int @@ -4395,6 +4395,12 @@ static bool intel_digital_port_connected(struct drm_i915_private *dev_priv, yesno(intel_dp_source_supports_hbr2(intel_dp)), yesno(drm_dp_tps3_supported(intel_dp->dpcd)));

• /* Set the max lane count for sink */

• intel_dp->max_sink_lane_count = drm_dp_max_lane_count(intel_dp->dpcd);

• /* Set the max link BW for sink */

• intel_dp->max_sink_link_bw = intel_dp_max_link_bw(intel_dp);

• intel_dp_print_rates(intel_dp);

  intel_dp_read_desc(intel_dp);

diff --git a/drivers/gpu/drm/i915/intel_drv.h b/drivers/gpu/drm/i915/intel_drv.h index fd77a3b..b6526ad 100644 --- a/drivers/gpu/drm/i915/intel_drv.h +++ b/drivers/gpu/drm/i915/intel_drv.h @@ -906,6 +906,10 @@ struct intel_dp { /* sink rates as reported by DP_SUPPORTED_LINK_RATES */ uint8_t num_sink_rates; int sink_rates[DP_MAX_SUPPORTED_RATES];

• /* Max lane count for the sink as per DPCD registers */
• uint8_t max_sink_lane_count;
• /* Max link BW for the sink as per DPCD registers */
• int max_sink_link_bw; /* sink or branch descriptor */ struct intel_dp_desc desc; struct drm_dp_aux aux;

-- 1.9.1

On Thu, Dec 08, 2016 at 11:23:39PM +0200, Jani Nikula wrote:

On Tue, 06 Dec 2016, Manasi Navare manasi.d.navare@intel.com wrote:

...

Sink's capabilities are advertised through DPCD registers and get updated only on hotplug. So they should be computed only once in the long pulse handler and saved off in intel_dp structure for the use later. For this reason two new fields max_sink_lane_count and max_sink_link_bw are added to intel_dp structure.

This also simplifies the fallback link rate/lane count logic to handle link training failure. In that case, the max_sink_link_bw and max_sink_lane_count can be reccomputed to match the fallback values lowering the sink capabilities due to link train failure.

Cc: Ville Syrjala ville.syrjala@linux.intel.com Cc: Jani Nikula jani.nikula@linux.intel.com Cc: Daniel Vetter daniel.vetter@intel.com Signed-off-by: Manasi Navare manasi.d.navare@intel.com

Reviewed-by: Jani Nikula jani.nikula@intel.com

Eventually we may want to call the fields *link* rates, because that's what they'll effectively be. Transient values that don't reflect the sink or source capabilities, but the link capabilities.

Thanks Jani for your r-b. Yes I agree, should I change the name to link_rate instead of sink_link_rate now? The only reason I kept it as sink_link_rate is because we go ahead and calculate the link capabilities later based on common link rate.

Regards Manasi

...

---

drivers/gpu/drm/i915/intel_dp.c | 10 ++++++++-- drivers/gpu/drm/i915/intel_drv.h | 4 ++++ 2 files changed, 12 insertions(+), 2 deletions(-)

diff --git a/drivers/gpu/drm/i915/intel_dp.c b/drivers/gpu/drm/i915/intel_dp.c index db75bb9..434dc7d 100644 --- a/drivers/gpu/drm/i915/intel_dp.c +++ b/drivers/gpu/drm/i915/intel_dp.c @@ -156,7 +156,7 @@ static u8 intel_dp_max_lane_count(struct intel_dp *intel_dp) u8 source_max, sink_max;

source_max = intel_dig_port->max_lanes;

• sink_max = drm_dp_max_lane_count(intel_dp->dpcd);

• sink_max = intel_dp->max_sink_lane_count;

  return min(source_max, sink_max);

} @@ -213,7 +213,7 @@ static u8 intel_dp_max_lane_count(struct intel_dp *intel_dp)

*sink_rates = default_rates;

• return (intel_dp_max_link_bw(intel_dp) >> 3) + 1;

• return (intel_dp->max_sink_link_bw >> 3) + 1;

}

static int @@ -4395,6 +4395,12 @@ static bool intel_digital_port_connected(struct drm_i915_private *dev_priv, yesno(intel_dp_source_supports_hbr2(intel_dp)), yesno(drm_dp_tps3_supported(intel_dp->dpcd)));

• /* Set the max lane count for sink */

• intel_dp->max_sink_lane_count = drm_dp_max_lane_count(intel_dp->dpcd);

• /* Set the max link BW for sink */

• intel_dp->max_sink_link_bw = intel_dp_max_link_bw(intel_dp);

• intel_dp_print_rates(intel_dp);

  intel_dp_read_desc(intel_dp);

diff --git a/drivers/gpu/drm/i915/intel_drv.h b/drivers/gpu/drm/i915/intel_drv.h index fd77a3b..b6526ad 100644 --- a/drivers/gpu/drm/i915/intel_drv.h +++ b/drivers/gpu/drm/i915/intel_drv.h @@ -906,6 +906,10 @@ struct intel_dp { /* sink rates as reported by DP_SUPPORTED_LINK_RATES */ uint8_t num_sink_rates; int sink_rates[DP_MAX_SUPPORTED_RATES];

• /* Max lane count for the sink as per DPCD registers */
• uint8_t max_sink_lane_count;
• /* Max link BW for the sink as per DPCD registers */
• int max_sink_link_bw; /* sink or branch descriptor */ struct intel_dp_desc desc; struct drm_dp_aux aux;

-- Jani Nikula, Intel Open Source Technology Center

On Thu, 08 Dec 2016, Manasi Navare manasi.d.navare@intel.com wrote:

Daniel, can we merge this patch?

Pushed this one to dinq, thanks for the patch.

BR, Jani.

It has no dependency on other link train patches, it is just a clean up for the existing driver code that uses max link rate and lane count values. Other link train patches have dependency on this thats why it was part of the series. But it would be great if this gets merged so that i dont have to rebase it tremendously after 3 weeks when I am back and plus its good for the driver to start using this clean up. This kind of clean up was long due.

Regards Manasi

On Thu, Dec 08, 2016 at 11:23:39PM +0200, Jani Nikula wrote:

...

On Tue, 06 Dec 2016, Manasi Navare manasi.d.navare@intel.com wrote:

...

Sink's capabilities are advertised through DPCD registers and get updated only on hotplug. So they should be computed only once in the long pulse handler and saved off in intel_dp structure for the use later. For this reason two new fields max_sink_lane_count and max_sink_link_bw are added to intel_dp structure.

This also simplifies the fallback link rate/lane count logic to handle link training failure. In that case, the max_sink_link_bw and max_sink_lane_count can be reccomputed to match the fallback values lowering the sink capabilities due to link train failure.

Cc: Ville Syrjala ville.syrjala@linux.intel.com Cc: Jani Nikula jani.nikula@linux.intel.com Cc: Daniel Vetter daniel.vetter@intel.com Signed-off-by: Manasi Navare manasi.d.navare@intel.com

Reviewed-by: Jani Nikula jani.nikula@intel.com

Eventually we may want to call the fields *link* rates, because that's what they'll effectively be. Transient values that don't reflect the sink or source capabilities, but the link capabilities.

...

---

drivers/gpu/drm/i915/intel_dp.c | 10 ++++++++-- drivers/gpu/drm/i915/intel_drv.h | 4 ++++ 2 files changed, 12 insertions(+), 2 deletions(-)

diff --git a/drivers/gpu/drm/i915/intel_dp.c b/drivers/gpu/drm/i915/intel_dp.c index db75bb9..434dc7d 100644 --- a/drivers/gpu/drm/i915/intel_dp.c +++ b/drivers/gpu/drm/i915/intel_dp.c @@ -156,7 +156,7 @@ static u8 intel_dp_max_lane_count(struct intel_dp *intel_dp) u8 source_max, sink_max;

source_max = intel_dig_port->max_lanes;

• sink_max = drm_dp_max_lane_count(intel_dp->dpcd);

• sink_max = intel_dp->max_sink_lane_count;

  return min(source_max, sink_max);

} @@ -213,7 +213,7 @@ static u8 intel_dp_max_lane_count(struct intel_dp *intel_dp)

*sink_rates = default_rates;

• return (intel_dp_max_link_bw(intel_dp) >> 3) + 1;

• return (intel_dp->max_sink_link_bw >> 3) + 1;

}

static int @@ -4395,6 +4395,12 @@ static bool intel_digital_port_connected(struct drm_i915_private *dev_priv, yesno(intel_dp_source_supports_hbr2(intel_dp)), yesno(drm_dp_tps3_supported(intel_dp->dpcd)));

• /* Set the max lane count for sink */

• intel_dp->max_sink_lane_count = drm_dp_max_lane_count(intel_dp->dpcd);

• /* Set the max link BW for sink */

• intel_dp->max_sink_link_bw = intel_dp_max_link_bw(intel_dp);

• intel_dp_print_rates(intel_dp);

  intel_dp_read_desc(intel_dp);

diff --git a/drivers/gpu/drm/i915/intel_drv.h b/drivers/gpu/drm/i915/intel_drv.h index fd77a3b..b6526ad 100644 --- a/drivers/gpu/drm/i915/intel_drv.h +++ b/drivers/gpu/drm/i915/intel_drv.h @@ -906,6 +906,10 @@ struct intel_dp { /* sink rates as reported by DP_SUPPORTED_LINK_RATES */ uint8_t num_sink_rates; int sink_rates[DP_MAX_SUPPORTED_RATES];

• /* Max lane count for the sink as per DPCD registers */
• uint8_t max_sink_lane_count;
• /* Max link BW for the sink as per DPCD registers */
• int max_sink_link_bw; /* sink or branch descriptor */ struct intel_dp_desc desc; struct drm_dp_aux aux;

-- Jani Nikula, Intel Open Source Technology Center

Hi Jani,

This patch uses the simplified fallback logic where if link train fails, we first lower the max sink link rate to a lower value until it reaches RBR keeping the max lane count unchanged and then lower the lane count and set max lane count to this lowered lane count and resetting the link rate to the max sink link rate based on DPCD values. Because this directly goes and updates the max link rate and max lane count values on fallback, we do not need any special case for fallback during intel_dp_compute_config.

Manasi

On Mon, Dec 05, 2016 at 04:27:37PM -0800, Manasi Navare wrote:

If link training fails, then we need to fallback to lower link rate first and if link training fails at RBR, then fallback to lower lane count. This function finds the next lower link rate/lane count value after link training failure and limits the max link_rate and lane_count values to these fallback values.

v6:

• Cap the max link rate and lane count to the max

values obtained during fallback link training (Daniel Vetter) v5:

• Start the fallback at the lane count value passed not

the max lane count (Jani Nikula) v4:

• Remove the redundant variable link_train_failed

v3:

• Remove fallback_link_rate_index variable, just obtain

that using the helper intel_dp_link_rate_index (Jani Nikula) v2: Squash the patch that returns the link rate index (Jani Nikula)

Acked-by: Tony Cheng tony.cheng@amd.com Acked-by: Harry Wentland harry.wentland@amd.com Cc: Ville Syrjala ville.syrjala@linux.intel.com Cc: Jani Nikula jani.nikula@linux.intel.com Cc: Daniel Vetter daniel.vetter@intel.com Signed-off-by: Manasi Navare manasi.d.navare@intel.com

---

drivers/gpu/drm/i915/intel_dp.c | 40 ++++++++++++++++++++++++++++++++++++++++ drivers/gpu/drm/i915/intel_drv.h | 2 ++ 2 files changed, 42 insertions(+)

diff --git a/drivers/gpu/drm/i915/intel_dp.c b/drivers/gpu/drm/i915/intel_dp.c index 434dc7d..b5c7526f 100644 --- a/drivers/gpu/drm/i915/intel_dp.c +++ b/drivers/gpu/drm/i915/intel_dp.c @@ -278,6 +278,46 @@ static int intel_dp_common_rates(struct intel_dp *intel_dp, common_rates); }

+static int intel_dp_link_rate_index(struct intel_dp *intel_dp,

• 		    int *common_rates, int link_rate)
  

+{

• int common_len;
• int index;
• common_len = intel_dp_common_rates(intel_dp, common_rates);
• for (index = 0; index < common_len; index++) {

• if (link_rate == common_rates[common_len - index - 1])
  

• 	return common_len - index - 1;
  

• }
• return -1;

+}

+int intel_dp_get_link_train_fallback_values(struct intel_dp *intel_dp,

• 			    int link_rate, uint8_t lane_count)
  

+{

• int common_rates[DP_MAX_SUPPORTED_RATES] = {};
• int common_len;
• int link_rate_index = -1;
• common_len = intel_dp_common_rates(intel_dp, common_rates);
• link_rate_index = intel_dp_link_rate_index(intel_dp,

• 				   common_rates,
  

• 				   link_rate);
  

• if (link_rate_index > 0) {

• intel_dp->max_sink_link_bw = drm_dp_link_rate_to_bw_code(common_rates[link_rate_index - 1]);
  

• intel_dp->max_sink_lane_count = lane_count;
  

• } else if (lane_count > 1) {

• intel_dp->max_sink_link_bw = intel_dp_max_link_bw(intel_dp);
  

• intel_dp->max_sink_lane_count = lane_count >> 1;
  

• } else {

• DRM_ERROR("Link Training Unsuccessful\n");
  

• return -1;
  

• }
• return 0;

+}

static enum drm_mode_status intel_dp_mode_valid(struct drm_connector *connector, struct drm_display_mode *mode) diff --git a/drivers/gpu/drm/i915/intel_drv.h b/drivers/gpu/drm/i915/intel_drv.h index b6526ad..47e3671 100644 --- a/drivers/gpu/drm/i915/intel_drv.h +++ b/drivers/gpu/drm/i915/intel_drv.h @@ -1400,6 +1400,8 @@ bool intel_dp_init_connector(struct intel_digital_port *intel_dig_port, void intel_dp_set_link_params(struct intel_dp *intel_dp, int link_rate, uint8_t lane_count, bool link_mst); +int intel_dp_get_link_train_fallback_values(struct intel_dp *intel_dp,

• 			    int link_rate, uint8_t lane_count);
  

void intel_dp_start_link_train(struct intel_dp *intel_dp); void intel_dp_stop_link_train(struct intel_dp *intel_dp); void intel_dp_sink_dpms(struct intel_dp *intel_dp, int mode); -- 1.9.1

On Thu, Dec 08, 2016 at 11:46:02PM +0200, Jani Nikula wrote:

On Tue, 06 Dec 2016, Manasi Navare manasi.d.navare@intel.com wrote:

...

If link training fails, then we need to fallback to lower link rate first and if link training fails at RBR, then fallback to lower lane count. This function finds the next lower link rate/lane count value after link training failure and limits the max link_rate and lane_count values to these fallback values.

v6:

• Cap the max link rate and lane count to the max

values obtained during fallback link training (Daniel Vetter) v5:

• Start the fallback at the lane count value passed not

the max lane count (Jani Nikula) v4:

• Remove the redundant variable link_train_failed

v3:

• Remove fallback_link_rate_index variable, just obtain

that using the helper intel_dp_link_rate_index (Jani Nikula) v2: Squash the patch that returns the link rate index (Jani Nikula)

Acked-by: Tony Cheng tony.cheng@amd.com Acked-by: Harry Wentland harry.wentland@amd.com Cc: Ville Syrjala ville.syrjala@linux.intel.com Cc: Jani Nikula jani.nikula@linux.intel.com Cc: Daniel Vetter daniel.vetter@intel.com Signed-off-by: Manasi Navare manasi.d.navare@intel.com

---

drivers/gpu/drm/i915/intel_dp.c | 40 ++++++++++++++++++++++++++++++++++++++++ drivers/gpu/drm/i915/intel_drv.h | 2 ++ 2 files changed, 42 insertions(+)

diff --git a/drivers/gpu/drm/i915/intel_dp.c b/drivers/gpu/drm/i915/intel_dp.c index 434dc7d..b5c7526f 100644 --- a/drivers/gpu/drm/i915/intel_dp.c +++ b/drivers/gpu/drm/i915/intel_dp.c @@ -278,6 +278,46 @@ static int intel_dp_common_rates(struct intel_dp *intel_dp, common_rates); }

+static int intel_dp_link_rate_index(struct intel_dp *intel_dp,

• 		    int *common_rates, int link_rate)
  

+{

• int common_len;
• int index;
• common_len = intel_dp_common_rates(intel_dp, common_rates);
• for (index = 0; index < common_len; index++) {

• if (link_rate == common_rates[common_len - index - 1])
  

• 	return common_len - index - 1;
  

Probably somewhere in the history of the patch series there was a time when it was necessary to search for the rates in reverse order. What possible benefit could that offer at this point?

The advantage here is that the link rate is more likely to match quicker if we search in reverse order.

...

• }
• return -1;

+}

+int intel_dp_get_link_train_fallback_values(struct intel_dp *intel_dp,

• 			    int link_rate, uint8_t lane_count)
  

+{

• int common_rates[DP_MAX_SUPPORTED_RATES] = {};

No need to initialize because you initialize it a couple of lines later.

Agreed

...

• int common_len;
• int link_rate_index = -1;

No need to initialize because you initialize it a couple of lines later.

...

• common_len = intel_dp_common_rates(intel_dp, common_rates);
• link_rate_index = intel_dp_link_rate_index(intel_dp,

• 				   common_rates,
  

• 				   link_rate);
  

Please stop and think, and don't rush each new iteration of the patches.

What's wrong with the above lines? Please think about it. Answer at the end of the mail (*).

...

• if (link_rate_index > 0) {

• intel_dp->max_sink_link_bw = drm_dp_link_rate_to_bw_code(common_rates[link_rate_index - 1]);
  

• intel_dp->max_sink_lane_count = lane_count;
  

• } else if (lane_count > 1) {

• intel_dp->max_sink_link_bw = intel_dp_max_link_bw(intel_dp);
  

• intel_dp->max_sink_lane_count = lane_count >> 1;
  

• } else {

• DRM_ERROR("Link Training Unsuccessful\n");
  

• return -1;
  

• }
• return 0;

+}

static enum drm_mode_status intel_dp_mode_valid(struct drm_connector *connector, struct drm_display_mode *mode) diff --git a/drivers/gpu/drm/i915/intel_drv.h b/drivers/gpu/drm/i915/intel_drv.h index b6526ad..47e3671 100644 --- a/drivers/gpu/drm/i915/intel_drv.h +++ b/drivers/gpu/drm/i915/intel_drv.h @@ -1400,6 +1400,8 @@ bool intel_dp_init_connector(struct intel_digital_port *intel_dig_port, void intel_dp_set_link_params(struct intel_dp *intel_dp, int link_rate, uint8_t lane_count, bool link_mst); +int intel_dp_get_link_train_fallback_values(struct intel_dp *intel_dp,

• 			    int link_rate, uint8_t lane_count);
  

void intel_dp_start_link_train(struct intel_dp *intel_dp); void intel_dp_stop_link_train(struct intel_dp *intel_dp); void intel_dp_sink_dpms(struct intel_dp *intel_dp, int mode);

(*) You do intel_dp_common_rates(intel_dp, common_rates) twice, for no reason at all.

Actually the first call was to obtain the common_len which was needed earlier but we no longer need it because of the simplified fallback logic modifying the max sink link rate directly. So yes I will remove the first call to intel_dp_common_rate() Good catch! Thanks Jani.

Regards Manasi

-- Jani Nikula, Intel Open Source Technology Center

On Fri, 09 Dec 2016, Manasi Navare manasi.d.navare@intel.com wrote:

If link training fails, then we need to fallback to lower link rate first and if link training fails at RBR, then fallback to lower lane count. This function finds the next lower link rate/lane count value after link training failure and limits the max link_rate and lane_count values to these fallback values.

v7:

• Remove unnecessary intializations and remove redundant

call to intel_dp_common_rates (Jani Nikula) v6:

• Cap the max link rate and lane count to the max

values obtained during fallback link training (Daniel Vetter) v5:

• Start the fallback at the lane count value passed not

the max lane count (Jani Nikula) v4:

• Remove the redundant variable link_train_failed

v3:

• Remove fallback_link_rate_index variable, just obtain

that using the helper intel_dp_link_rate_index (Jani Nikula) v2: Squash the patch that returns the link rate index (Jani Nikula)

Acked-by: Tony Cheng tony.cheng@amd.com Acked-by: Harry Wentland harry.wentland@amd.com Cc: Ville Syrjala ville.syrjala@linux.intel.com Cc: Jani Nikula jani.nikula@linux.intel.com Cc: Daniel Vetter daniel.vetter@intel.com Signed-off-by: Manasi Navare manasi.d.navare@intel.com

Reviewed-by: Jani Nikula jani.nikula@intel.com

---

drivers/gpu/drm/i915/intel_dp.c | 38 ++++++++++++++++++++++++++++++++++++++ drivers/gpu/drm/i915/intel_drv.h | 2 ++ 2 files changed, 40 insertions(+)

diff --git a/drivers/gpu/drm/i915/intel_dp.c b/drivers/gpu/drm/i915/intel_dp.c index 434dc7d..1b5e0d0 100644 --- a/drivers/gpu/drm/i915/intel_dp.c +++ b/drivers/gpu/drm/i915/intel_dp.c @@ -278,6 +278,44 @@ static int intel_dp_common_rates(struct intel_dp *intel_dp, common_rates); }

+static int intel_dp_link_rate_index(struct intel_dp *intel_dp,

• 		    int *common_rates, int link_rate)
  

+{

• int common_len;
• int index;
• common_len = intel_dp_common_rates(intel_dp, common_rates);
• for (index = 0; index < common_len; index++) {

• if (link_rate == common_rates[common_len - index - 1])
  

• 	return common_len - index - 1;
  

• }
• return -1;

+}

+int intel_dp_get_link_train_fallback_values(struct intel_dp *intel_dp,

• 			    int link_rate, uint8_t lane_count)
  

+{

• int common_rates[DP_MAX_SUPPORTED_RATES];
• int link_rate_index;
• link_rate_index = intel_dp_link_rate_index(intel_dp,

• 				   common_rates,
  

• 				   link_rate);
  

• if (link_rate_index > 0) {

• intel_dp->max_sink_link_bw = drm_dp_link_rate_to_bw_code(common_rates[link_rate_index - 1]);
  

• intel_dp->max_sink_lane_count = lane_count;
  

• } else if (lane_count > 1) {

• intel_dp->max_sink_link_bw = intel_dp_max_link_bw(intel_dp);
  

• intel_dp->max_sink_lane_count = lane_count >> 1;
  

• } else {

• DRM_ERROR("Link Training Unsuccessful\n");
  

• return -1;
  

• }
• return 0;

+}

static enum drm_mode_status intel_dp_mode_valid(struct drm_connector *connector, struct drm_display_mode *mode) diff --git a/drivers/gpu/drm/i915/intel_drv.h b/drivers/gpu/drm/i915/intel_drv.h index 377693e..0a20422 100644 --- a/drivers/gpu/drm/i915/intel_drv.h +++ b/drivers/gpu/drm/i915/intel_drv.h @@ -1400,6 +1400,8 @@ bool intel_dp_init_connector(struct intel_digital_port *intel_dig_port, void intel_dp_set_link_params(struct intel_dp *intel_dp, int link_rate, uint8_t lane_count, bool link_mst); +int intel_dp_get_link_train_fallback_values(struct intel_dp *intel_dp,

• 			    int link_rate, uint8_t lane_count);
  

void intel_dp_start_link_train(struct intel_dp *intel_dp); void intel_dp_stop_link_train(struct intel_dp *intel_dp); void intel_dp_sink_dpms(struct intel_dp *intel_dp, int mode);

On Tue, 06 Dec 2016, Manasi Navare manasi.d.navare@intel.com wrote:

If link training at a link rate optimal for a particular mode fails during modeset's atomic commit phase, then we let the modeset complete and then retry. We save the link rate value at which link training failed, update the link status property to "BAD" and use a lower link rate to prune the modes. It will redo the modeset on the current mode at lower link rate or if the current mode gets pruned due to lower link constraints then, it will send a hotplug uevent for userspace to handle it.

This is also required to pass DP CTS tests 4.3.1.3, 4.3.1.4, 4.3.1.6.

v9:

• Use the trimmed max values of link rate/lane count based on

link train fallback (Daniel Vetter) v8:

• Set link_status to BAD first and then call mode_valid (Jani Nikula)

v7: Remove the redundant variable in previous patch itself v6:

• Obtain link rate index from fallback_link_rate using

the helper intel_dp_link_rate_index (Jani Nikula)

• Include fallback within intel_dp_start_link_train (Jani Nikula)

v5:

• Move set link status to drm core (Daniel Vetter, Jani Nikula)

v4:

• Add fallback support for non DDI platforms too
• Set connector->link status inside set_link_status function

(Jani Nikula) v3:

• Set link status property to BAd unconditionally (Jani Nikula)
• Dont use two separate variables link_train_failed and link_status

to indicate same thing (Jani Nikula) v2:

• Squashed a few patches (Jani Nikula)

Acked-by: Tony Cheng tony.cheng@amd.com Acked-by: Harry Wentland Harry.wentland@amd.com Cc: Jani Nikula jani.nikula@linux.intel.com Cc: Daniel Vetter daniel.vetter@intel.com Cc: Ville Syrjala ville.syrjala@linux.intel.com Signed-off-by: Manasi Navare manasi.d.navare@intel.com

This is not a journey I'd like to do again, but I like the place where we've finally arrived.

Reviewed-by: Jani Nikula jani.nikula@intel.com

---

drivers/gpu/drm/i915/intel_dp.c | 27 +++++++++++++++++++++++++++ drivers/gpu/drm/i915/intel_dp_link_training.c | 22 ++++++++++++++++++++-- drivers/gpu/drm/i915/intel_drv.h | 3 +++ 3 files changed, 50 insertions(+), 2 deletions(-)

diff --git a/drivers/gpu/drm/i915/intel_dp.c b/drivers/gpu/drm/i915/intel_dp.c index b5c7526f..404da32 100644 --- a/drivers/gpu/drm/i915/intel_dp.c +++ b/drivers/gpu/drm/i915/intel_dp.c @@ -5676,6 +5676,29 @@ static bool intel_edp_init_connector(struct intel_dp *intel_dp, return false; }

+static void intel_dp_modeset_retry_work_fn(struct work_struct *work) +{

• struct intel_connector *intel_connector;
• struct drm_connector *connector;
• intel_connector = container_of(work, typeof(*intel_connector),

• 		       modeset_retry_work);
  

• connector = &intel_connector->base;
• DRM_DEBUG_KMS("[CONNECTOR:%d:%s]\n", connector->base.id,

•       connector->name);
  

• /* Grab the locks before changing connector property*/
• mutex_lock(&connector->dev->mode_config.mutex);
• /* Set connector link status to BAD and send a Uevent to notify

• * userspace to do a modeset.
  

• */
  

• drm_mode_connector_set_link_status_property(connector,

• 				    DRM_MODE_LINK_STATUS_BAD);
  

• mutex_unlock(&connector->dev->mode_config.mutex);
• /* Send Hotplug uevent so userspace can reprobe */
• drm_kms_helper_hotplug_event(connector->dev);

+}

bool intel_dp_init_connector(struct intel_digital_port *intel_dig_port, struct intel_connector *intel_connector) @@ -5688,6 +5711,10 @@ static bool intel_edp_init_connector(struct intel_dp *intel_dp, enum port port = intel_dig_port->port; int type;

• /* Initialize the work for modeset in case of link train failure */
• INIT_WORK(&intel_connector->modeset_retry_work,

•   intel_dp_modeset_retry_work_fn);
  

• if (WARN(intel_dig_port->max_lanes < 1, "Not enough lanes (%d) for DP on port %c\n", intel_dig_port->max_lanes, port_name(port)))

diff --git a/drivers/gpu/drm/i915/intel_dp_link_training.c b/drivers/gpu/drm/i915/intel_dp_link_training.c index 0048b52..955b239 100644 --- a/drivers/gpu/drm/i915/intel_dp_link_training.c +++ b/drivers/gpu/drm/i915/intel_dp_link_training.c @@ -313,6 +313,24 @@ void intel_dp_stop_link_train(struct intel_dp *intel_dp) void intel_dp_start_link_train(struct intel_dp *intel_dp) {

• intel_dp_link_training_clock_recovery(intel_dp);
• intel_dp_link_training_channel_equalization(intel_dp);

• struct intel_connector *intel_connector = intel_dp->attached_connector;
• if (!intel_dp_link_training_clock_recovery(intel_dp))

• goto failure_handling;
  

• if (!intel_dp_link_training_channel_equalization(intel_dp))

• goto failure_handling;
  

• DRM_DEBUG_KMS("Link Training Passed at Link Rate = %d, Lane count = %d",

•       intel_dp->link_rate, intel_dp->lane_count);
  

• return;
• failure_handling:
• DRM_DEBUG_KMS("Link Training failed at link rate = %d, lane count = %d",

•       intel_dp->link_rate, intel_dp->lane_count);
  

• if (!intel_dp_get_link_train_fallback_values(intel_dp,

• 				     intel_dp->link_rate,
  

• 				     intel_dp->lane_count))
  

• /* Schedule a Hotplug Uevent to userspace to start modeset */
  

• schedule_work(&intel_connector->modeset_retry_work);
  

• return;

} diff --git a/drivers/gpu/drm/i915/intel_drv.h b/drivers/gpu/drm/i915/intel_drv.h index 47e3671..9d6f7ee 100644 --- a/drivers/gpu/drm/i915/intel_drv.h +++ b/drivers/gpu/drm/i915/intel_drv.h @@ -315,6 +315,9 @@ struct intel_connector { void *port; /* store this opaque as its illegal to dereference it */

struct intel_dp *mst_port;

• /* Work struct to schedule a uevent on link train failure */
• struct work_struct modeset_retry_work;

};

struct dpll {