Skip to contentSkip to conversation

drm/i915: Implement Link Rate fallback on Link training failure

PATCH In review Manasi Navare 3 participants 7 messages Mar 14, 2017 – Mar 31, 2017

Subsystems (2)
lore
Share & tools
lore
#0 lore raw

PATCH v10 1/1 drm/i915: Implement Link Rate fallback on Link training failure

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.

This patch is a resend of the original commit id (233ce881dd91fb "drm/i915: Implement Link Rate fallback on Link training failure") which got reverted in this commit id (afc1ebf4562a14 Revert "drm/i915: Implement Link Rate fallback on Link training failure") due to CI failures.

After investigating the CI failures it was found that these were essentially the failures which were always there but hidden because they used to be DRM_DEBUG_KMS messages for link failures so never got caught by CI. But now this patch actually throws DRM_ERROR if the link training fails at RBR and 1 lane. So it caught these link train failures.

There were two failures:
1. On SKL 6700k this was because the machine in CI lab is a SKL desktop without eDP on Port A. But our VBT initialization code in the driver writes VBT defaults in a way that it always sets DP flag on Port A and this does not get cleared after parsing the VBT outputs. This has been fixed in commit id (bb1d132935c2f8 "drm/i915/vbt: split out defaults that are set when there is no VBT) and (665788572c6410b "drm/i915/vbt: don't propagate errors from intel_bios_init())

2. On ILK-650 desktop - This was happening because of a bad monitor desktop combination. I switched the monitor in the CI lab and that helped get rid of the link failures on ILK system.

v10:
* Rebase on drm-tip and resend after revert
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)

---
 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(-)
drivers/gpu/drm/i915/intel_dp.c +27 -0
index fd96a6c..895f934 100644
@@ -5924,6 +5924,29 @@ intel_dp_init_connector_port_info(struct intel_digital_port *intel_dig_port)
 	}
 }
 
+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)
@@ -5936,6 +5959,10 @@ intel_dp_init_connector(struct intel_digital_port *intel_dig_port,
 	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)))
drivers/gpu/drm/i915/intel_dp_link_training.c +20 -2
index 0048b52..955b239 100644
@@ -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;
 }
drivers/gpu/drm/i915/intel_drv.h +3 -0
index 51228fe..0fe1ac8 100644
@@ -321,6 +321,9 @@ 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 {
Signature · 6 lines
2.1.4

_______________________________________________
Intel-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/intel-gfx
#1 lore raw

✗ Fi.CI.BAT: failure for drm/i915: Implement Link Rate fallback on Link training failure

== Series Details ==

== Series Details ==

Series: drm/i915: Implement Link Rate fallback on Link training failure
URL : https://patchwork.freedesktop.org/series/21248/
State : failure

== Summary ==

Series 21248v1 drm/i915: Implement Link Rate fallback on Link training failure
https://patchwork.freedesktop.org/api/1.0/series/21248/revisions/1/mbox/

Test gem_exec_flush:

        Subgroup basic-batch-kernel-default-uc:
                pass       -> FAIL       (fi-snb-2600) fdo#100007
        Subgroup basic-uc-pro-default:
                pass       -> INCOMPLETE (fi-skl-6770hq) fdo#100130
        Subgroup basic-uc-ro-default:
                pass       -> INCOMPLETE (fi-skl-6700hq) fdo#100130
        Subgroup basic-uc-rw-default:
                pass       -> INCOMPLETE (fi-skl-6260u)
Test kms_cursor_legacy:
        Subgroup basic-busy-flip-before-cursor-atomic:
                fail       -> PASS       (fi-snb-2520m)

fdo#100007 https://bugs.freedesktop.org/show_bug.cgi?id=100007
fdo#100130 https://bugs.freedesktop.org/show_bug.cgi?id=100130

fi-bdw-5557u     total:278  pass:267  dwarn:0   dfail:0   fail:0   skip:11  time: 459s
fi-bsw-n3050     total:278  pass:239  dwarn:0   dfail:0   fail:0   skip:39  time: 575s
fi-bxt-j4205     total:278  pass:259  dwarn:0   dfail:0   fail:0   skip:19  time: 538s
fi-bxt-t5700     total:278  pass:258  dwarn:0   dfail:0   fail:0   skip:20  time: 559s
fi-byt-j1900     total:278  pass:251  dwarn:0   dfail:0   fail:0   skip:27  time: 503s
fi-byt-n2820     total:278  pass:247  dwarn:0   dfail:0   fail:0   skip:31  time: 492s
fi-ilk-650       total:278  pass:228  dwarn:0   dfail:0   fail:0   skip:50  time: 430s
fi-ivb-3770      total:278  pass:260  dwarn:0   dfail:0   fail:0   skip:18  time: 501s
fi-skl-6260u     total:56   pass:54   dwarn:0   dfail:0   fail:0   skip:1   time: 0s
fi-skl-6700hq    total:55   pass:47   dwarn:0   dfail:0   fail:0   skip:7   time: 0s
fi-skl-6770hq    total:53   pass:51   dwarn:0   dfail:0   fail:0   skip:1   time: 0s
fi-snb-2520m     total:278  pass:250  dwarn:0   dfail:0   fail:0   skip:28  time: 548s
fi-snb-2600      total:278  pass:248  dwarn:0   dfail:0   fail:1   skip:29  time: 416s

c641417b70c6b78efca29ae732d7cbf5716ac6d5 drm-tip: 2017y-03m-14d-16h-04m-56s UTC integration manifest 68a2710 drm/i915: Implement Link Rate fallback on Link training failure

== Logs ==

For more details see: https://intel-gfx-ci.01.org/CI/Patchwork_4173/
_______________________________________________
Intel-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/intel-gfx

#2 lore raw

✓ Fi.CI.BAT: success for drm/i915: Implement Link Rate fallback on Link training failure

In reply to Manasi Navare
== Series Details ==

== Series Details ==

Series: drm/i915: Implement Link Rate fallback on Link training failure
URL : https://patchwork.freedesktop.org/series/21248/
State : success

== Summary ==

Series 21248v1 drm/i915: Implement Link Rate fallback on Link training failure
https://patchwork.freedesktop.org/api/1.0/series/21248/revisions/1/mbox/

Test gem_exec_flush:

        Subgroup basic-batch-kernel-default-uc:
                pass       -> FAIL       (fi-snb-2600) fdo#100007
        Subgroup basic-uc-pro-default:
                pass       -> INCOMPLETE (fi-skl-6770hq) fdo#100130
        Subgroup basic-uc-ro-default:
                pass       -> INCOMPLETE (fi-skl-6700hq) fdo#100130
        Subgroup basic-uc-rw-default:
                pass       -> INCOMPLETE (fi-skl-6260u) fdo#100130
Test kms_cursor_legacy:
        Subgroup basic-busy-flip-before-cursor-atomic:
                fail       -> PASS       (fi-snb-2520m)

fdo#100007 https://bugs.freedesktop.org/show_bug.cgi?id=100007
fdo#100130 https://bugs.freedesktop.org/show_bug.cgi?id=100130

fi-bdw-5557u     total:278  pass:267  dwarn:0   dfail:0   fail:0   skip:11  time: 459s
fi-bsw-n3050     total:278  pass:239  dwarn:0   dfail:0   fail:0   skip:39  time: 575s
fi-bxt-j4205     total:278  pass:259  dwarn:0   dfail:0   fail:0   skip:19  time: 538s
fi-bxt-t5700     total:278  pass:258  dwarn:0   dfail:0   fail:0   skip:20  time: 559s
fi-byt-j1900     total:278  pass:251  dwarn:0   dfail:0   fail:0   skip:27  time: 503s
fi-byt-n2820     total:278  pass:247  dwarn:0   dfail:0   fail:0   skip:31  time: 492s
fi-ilk-650       total:278  pass:228  dwarn:0   dfail:0   fail:0   skip:50  time: 430s
fi-ivb-3770      total:278  pass:260  dwarn:0   dfail:0   fail:0   skip:18  time: 501s
fi-skl-6260u     total:56   pass:54   dwarn:0   dfail:0   fail:0   skip:1   time: 0s
fi-skl-6700hq    total:55   pass:47   dwarn:0   dfail:0   fail:0   skip:7   time: 0s
fi-skl-6770hq    total:53   pass:51   dwarn:0   dfail:0   fail:0   skip:1   time: 0s
fi-snb-2520m     total:278  pass:250  dwarn:0   dfail:0   fail:0   skip:28  time: 548s
fi-snb-2600      total:278  pass:248  dwarn:0   dfail:0   fail:1   skip:29  time: 416s

c641417b70c6b78efca29ae732d7cbf5716ac6d5 drm-tip: 2017y-03m-14d-16h-04m-56s UTC integration manifest 68a2710 drm/i915: Implement Link Rate fallback on Link training failure

== Logs ==

For more details see: https://intel-gfx-ci.01.org/CI/Patchwork_4173/
_______________________________________________
Intel-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/intel-gfx

#3 lore raw
In reply to Manasi Navare

Hi Jani/Ville,

I need to add another quick fix which would be required for the fallback to happen as expected. Should I respin this patch to add that fix or should I wait for this to get landed?

I have mentioned the fix suggested below, please let me know your thoughts on that.

Quoted 141 lines · Manasi Navare

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.

This patch is a resend of the original commit id (233ce881dd91fb "drm/i915: Implement Link Rate fallback on Link training failure") which got reverted in this commit id (afc1ebf4562a14 Revert "drm/i915: Implement Link Rate fallback on Link training failure") due to CI failures.

After investigating the CI failures it was found that these were essentially the failures which were always there but hidden because they used to be DRM_DEBUG_KMS messages for link failures so never got caught by CI. But now this patch actually throws DRM_ERROR if the link training fails at RBR and 1 lane. So it caught these link train failures.

There were two failures:
1. On SKL 6700k this was because the machine in CI lab is a SKL desktop without eDP on Port A. But our VBT initialization code in the driver writes VBT defaults in a way that it always sets DP flag on Port A and this does not get cleared after parsing the VBT outputs. This has been fixed in commit id (bb1d132935c2f8 "drm/i915/vbt: split out defaults that are set when there is no VBT) and (665788572c6410b "drm/i915/vbt: don't propagate errors from intel_bios_init())

2. On ILK-650 desktop - This was happening because of a bad monitor desktop combination. I switched the monitor in the CI lab and that helped get rid of the link failures on ILK system.

v10:
* Rebase on drm-tip and resend after revert
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)

---
 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(-)
drivers/gpu/drm/i915/intel_dp.c +27 -0
index fd96a6c..895f934 100644
@@ -5924,6 +5924,29 @@ intel_dp_init_connector_port_info(struct intel_digital_port *intel_dig_port)
 	}
 }
 
+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)
@@ -5936,6 +5959,10 @@ intel_dp_init_connector(struct intel_digital_port *intel_dig_port,
 	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)))
drivers/gpu/drm/i915/intel_dp_link_training.c +9 -2
index 0048b52..955b239 100644
@@ -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);

drivers/gpu/drm/i915/intel_dp_link_training.c void intel_dp_stop_link_train(struct intel_dp *intel_dp) · line 334

Quoted 10 lines
drivers/gpu/drm/i915/intel_dp_link_training.c +10 -0
+	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);

This is where the new boolean intel_dp->train_set_valid will have to be set to false on failure. This ensures that we dont end up retraining at the same parameters in intel_dp_check_link_status()

So in intel_dp_check-link_status , just before retraining it will add a check, if (!intel_dp->train_set_valid)
return;

Ville/Jani what are your thoughts on this?

Regards
Manasi

Quoted 19 lines
drivers/gpu/drm/i915/intel_dp_link_training.c +1 -0
+	return;
 }
drivers/gpu/drm/i915/intel_drv.h +3 -0
index 51228fe..0fe1ac8 100644
@@ -321,6 +321,9 @@ 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 {
Signature · 1 line
2.1.4

_______________________________________________
Intel-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/intel-gfx

#4 lore raw

Hi Jani,

I have reviewed your recent DP link rate / lane count refactoring patch series and I believe that it will soon get merged.

Are you waiting on merging this patch so that your refactoring series gets merged first? Or can we merge this patch since it already has ACKs and R-bs.

Please advise on next steps for this patch, since some of the PSR related patches are pending merge because of this.

Regards
Manasi

Quoted 189 lines · Manasi Navare

Hi Jani/Ville,

I need to add another quick fix which would be required for the fallback to happen as expected. Should I respin this patch to add that fix or should I wait for this to get landed?

I have mentioned the fix suggested below, please let me know your thoughts on that.

Quoted 141 lines · Manasi Navare

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.

This patch is a resend of the original commit id (233ce881dd91fb "drm/i915: Implement Link Rate fallback on Link training failure") which got reverted in this commit id (afc1ebf4562a14 Revert "drm/i915: Implement Link Rate fallback on Link training failure") due to CI failures.

After investigating the CI failures it was found that these were essentially the failures which were always there but hidden because they used to be DRM_DEBUG_KMS messages for link failures so never got caught by CI. But now this patch actually throws DRM_ERROR if the link training fails at RBR and 1 lane. So it caught these link train failures.

There were two failures:
1. On SKL 6700k this was because the machine in CI lab is a SKL desktop without eDP on Port A. But our VBT initialization code in the driver writes VBT defaults in a way that it always sets DP flag on Port A and this does not get cleared after parsing the VBT outputs. This has been fixed in commit id (bb1d132935c2f8 "drm/i915/vbt: split out defaults that are set when there is no VBT) and (665788572c6410b "drm/i915/vbt: don't propagate errors from intel_bios_init())

2. On ILK-650 desktop - This was happening because of a bad monitor desktop combination. I switched the monitor in the CI lab and that helped get rid of the link failures on ILK system.

v10:
* Rebase on drm-tip and resend after revert
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)

---
 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(-)
drivers/gpu/drm/i915/intel_dp.c +27 -0
index fd96a6c..895f934 100644
@@ -5924,6 +5924,29 @@ intel_dp_init_connector_port_info(struct intel_digital_port *intel_dig_port)
 	}
 }
 
+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)
@@ -5936,6 +5959,10 @@ intel_dp_init_connector(struct intel_digital_port *intel_dig_port,
 	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)))
drivers/gpu/drm/i915/intel_dp_link_training.c +9 -2
index 0048b52..955b239 100644
@@ -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);

drivers/gpu/drm/i915/intel_dp_link_training.c void intel_dp_stop_link_train(struct intel_dp *intel_dp) · line 334

Quoted 10 lines
drivers/gpu/drm/i915/intel_dp_link_training.c +10 -0
+	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);

This is where the new boolean intel_dp->train_set_valid will have to be set to false on failure. This ensures that we dont end up retraining at the same parameters in intel_dp_check_link_status()

So in intel_dp_check-link_status , just before retraining it will add a check, if (!intel_dp->train_set_valid)
return;

Ville/Jani what are your thoughts on this?

Regards
Manasi

Quoted 15 lines
drivers/gpu/drm/i915/intel_dp_link_training.c +1 -0
+	return;
 }
drivers/gpu/drm/i915/intel_drv.h +3 -0
index 51228fe..0fe1ac8 100644
@@ -321,6 +321,9 @@ 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;
 };
 
Quoted 8 lines
Quoted 4 lines
drivers/gpu/drm/i915/intel_drv.h +0 -0
 struct dpll {
Signature · 1 line
2.1.4

_______________________________________________
Intel-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/intel-gfx

_______________________________________________
Intel-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/intel-gfx

#5 lore raw
Quoted 4 lines · Manasi Navare

Hi Jani,

I have reviewed your recent DP link rate / lane count refactoring patch series and I believe that it will soon get merged.

I haven't received reviews on all of the patches in my series yet, in particular the first patches are missing reviews so I can't even get started pushing them.

BR,
Jani.

Quoted 210 lines

Are you waiting on merging this patch so that your refactoring series gets merged first? Or can we merge this patch since it already has ACKs and R-bs.

Please advise on next steps for this patch, since some of the PSR related patches are pending merge because of this.

Regards
Manasi

Quoted 197 lines · Manasi Navare

Hi Jani/Ville,

I need to add another quick fix which would be required for the fallback to happen as expected. Should I respin this patch to add that fix or should I wait for this to get landed?

I have mentioned the fix suggested below, please let me know your thoughts on that.

Quoted 141 lines · Manasi Navare

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.

This patch is a resend of the original commit id (233ce881dd91fb "drm/i915: Implement Link Rate fallback on Link training failure") which got reverted in this commit id (afc1ebf4562a14 Revert "drm/i915: Implement Link Rate fallback on Link training failure") due to CI failures.

After investigating the CI failures it was found that these were essentially the failures which were always there but hidden because they used to be DRM_DEBUG_KMS messages for link failures so never got caught by CI. But now this patch actually throws DRM_ERROR if the link training fails at RBR and 1 lane. So it caught these link train failures.

There were two failures:
1. On SKL 6700k this was because the machine in CI lab is a SKL desktop without eDP on Port A. But our VBT initialization code in the driver writes VBT defaults in a way that it always sets DP flag on Port A and this does not get cleared after parsing the VBT outputs. This has been fixed in commit id (bb1d132935c2f8 "drm/i915/vbt: split out defaults that are set when there is no VBT) and (665788572c6410b "drm/i915/vbt: don't propagate errors from intel_bios_init())

2. On ILK-650 desktop - This was happening because of a bad monitor desktop combination. I switched the monitor in the CI lab and that helped get rid of the link failures on ILK system.

v10:
* Rebase on drm-tip and resend after revert
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)

---
 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(-)
drivers/gpu/drm/i915/intel_dp.c +27 -0
index fd96a6c..895f934 100644
@@ -5924,6 +5924,29 @@ intel_dp_init_connector_port_info(struct intel_digital_port *intel_dig_port)
 	}
 }
 
+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)
@@ -5936,6 +5959,10 @@ intel_dp_init_connector(struct intel_digital_port *intel_dig_port,
 	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)))
drivers/gpu/drm/i915/intel_dp_link_training.c +9 -2
index 0048b52..955b239 100644
@@ -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);

drivers/gpu/drm/i915/intel_dp_link_training.c void intel_dp_stop_link_train(struct intel_dp *intel_dp) · line 334

Quoted 10 lines
drivers/gpu/drm/i915/intel_dp_link_training.c +10 -0
+	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);

This is where the new boolean intel_dp->train_set_valid will have to be set to false on failure. This ensures that we dont end up retraining at the same parameters in intel_dp_check_link_status()

So in intel_dp_check-link_status , just before retraining it will add a check, if (!intel_dp->train_set_valid)
return;

Ville/Jani what are your thoughts on this?

Regards
Manasi

Quoted 19 lines
drivers/gpu/drm/i915/intel_dp_link_training.c +1 -0
+	return;
 }
drivers/gpu/drm/i915/intel_drv.h +3 -0
index 51228fe..0fe1ac8 100644
@@ -321,6 +321,9 @@ 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 {
Signature · 1 line
2.1.4

_______________________________________________
Intel-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/intel-gfx

Signature · 5 lines
Jani Nikula, Intel Open Source Technology Center
_______________________________________________
Intel-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/intel-gfx
#6 lore raw
Quoted 6 lines · Jani Nikula
Quoted 4 lines · Manasi Navare

Hi Jani,

I have reviewed your recent DP link rate / lane count refactoring patch series and I believe that it will soon get merged.

Quoted 7 lines

I haven't received reviews on all of the patches in my series yet, in particular the first patches are missing reviews so I can't even get started pushing them.

BR,
Jani.

Ok, I will spend some time reviewing these patches today so you can push those.
Is that why you are waiting to merge this patch?

Regards
Manasi

Quoted 203 lines
Quoted 203 lines

Are you waiting on merging this patch so that your refactoring series gets merged first? Or can we merge this patch since it already has ACKs and R-bs.

Please advise on next steps for this patch, since some of the PSR related patches are pending merge because of this.

Regards
Manasi

Quoted 189 lines · Manasi Navare

Hi Jani/Ville,

I need to add another quick fix which would be required for the fallback to happen as expected. Should I respin this patch to add that fix or should I wait for this to get landed?

I have mentioned the fix suggested below, please let me know your thoughts on that.

Quoted 141 lines · Manasi Navare

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.

This patch is a resend of the original commit id (233ce881dd91fb "drm/i915: Implement Link Rate fallback on Link training failure") which got reverted in this commit id (afc1ebf4562a14 Revert "drm/i915: Implement Link Rate fallback on Link training failure") due to CI failures.

After investigating the CI failures it was found that these were essentially the failures which were always there but hidden because they used to be DRM_DEBUG_KMS messages for link failures so never got caught by CI. But now this patch actually throws DRM_ERROR if the link training fails at RBR and 1 lane. So it caught these link train failures.

There were two failures:
1. On SKL 6700k this was because the machine in CI lab is a SKL desktop without eDP on Port A. But our VBT initialization code in the driver writes VBT defaults in a way that it always sets DP flag on Port A and this does not get cleared after parsing the VBT outputs. This has been fixed in commit id (bb1d132935c2f8 "drm/i915/vbt: split out defaults that are set when there is no VBT) and (665788572c6410b "drm/i915/vbt: don't propagate errors from intel_bios_init())

2. On ILK-650 desktop - This was happening because of a bad monitor desktop combination. I switched the monitor in the CI lab and that helped get rid of the link failures on ILK system.

v10:
* Rebase on drm-tip and resend after revert
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)

---
 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(-)
drivers/gpu/drm/i915/intel_dp.c +27 -0
index fd96a6c..895f934 100644
@@ -5924,6 +5924,29 @@ intel_dp_init_connector_port_info(struct intel_digital_port *intel_dig_port)
 	}
 }
 
+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)
@@ -5936,6 +5959,10 @@ intel_dp_init_connector(struct intel_digital_port *intel_dig_port,
 	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)))
drivers/gpu/drm/i915/intel_dp_link_training.c +9 -2
index 0048b52..955b239 100644
@@ -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);

drivers/gpu/drm/i915/intel_dp_link_training.c void intel_dp_stop_link_train(struct intel_dp *intel_dp) · line 334

Quoted 10 lines
drivers/gpu/drm/i915/intel_dp_link_training.c +10 -0
+	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);

This is where the new boolean intel_dp->train_set_valid will have to be set to false on failure. This ensures that we dont end up retraining at the same parameters in intel_dp_check_link_status()

So in intel_dp_check-link_status , just before retraining it will add a check, if (!intel_dp->train_set_valid)
return;

Ville/Jani what are your thoughts on this?

Regards
Manasi

Quoted 15 lines
drivers/gpu/drm/i915/intel_dp_link_training.c +1 -0
+	return;
 }
drivers/gpu/drm/i915/intel_drv.h +3 -0
index 51228fe..0fe1ac8 100644
@@ -321,6 +321,9 @@ 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;
 };
 
Quoted 11 lines
Quoted 8 lines
Quoted 8 lines
Quoted 4 lines
drivers/gpu/drm/i915/intel_drv.h +0 -0
 struct dpll {
Signature · 1 line
2.1.4

_______________________________________________
Intel-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/intel-gfx

Signature · 1 line
Jani Nikula, Intel Open Source Technology Center

_______________________________________________
Intel-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/intel-gfx