Skip to contentSkip to conversation

HDMI optimization series

PATCH Applied Sonika Jindal 8 participants 74 messages Sep 4, 2015 – Mar 10, 2016

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

cover HDMI optimization series

This series adds changes in HDMI detection methods and also afew optimization. The overview of changes are:
1. HDMI EDID is read only at the hot-plug time.
2. EDID is cached in connectoer on hotplug,and released from cache only on the hot-unplug.
3. In between, for all detetct calls, only cached EDID is used.
4. HDMI EDID is read, only when live status is up.
5. HDMI is force probed during the connector init time, for connected boot scenarios.

v2:
Some refactoring is with this series.

Also, right now this is done for platforms gen7 and above because we couldn't test with older platforms. For newer platforms it works reliably.

For HPD and live status to work on SKL, following patch is required:
"drm/i915: Handle HPD when it has actually occurred"

v3:
Added retrial for live_status.
Relying on HPD fpr edid detection onlyfrom gen 8 onwards and VLV

v4:
*Also, including durga's patch for fixing irq_port for edp. Without this fix, hdmi hpd breaks with hdmi and edp connected on bxt. *Taking off the check for older platforms and now rely on init and hotplug to get the edid updated
*Some other cleanups.

Durgadoss R (1):
drm/i915/bxt: Fix irq_port for eDP

Shashank Sharma (2):
drm/i915: add attached connector to hdmi container drm/i915: Add HDMI probe function

Sonika Jindal (3):
drm/i915: Make intel_digital_port_connected global drm/i915: drm/i915: Check live status before reading edid drm/i915: drm/i915: Process hpd only for hdmi inside
hotplug_work_func

 drivers/gpu/drm/i915/intel_ddi.c     |   15 ++++---
 drivers/gpu/drm/i915/intel_dp.c      |   11 +++--
 drivers/gpu/drm/i915/intel_drv.h     |    3 ++
 drivers/gpu/drm/i915/intel_hdmi.c    |   80 +++++++++++++++++++++++++++++-----
 drivers/gpu/drm/i915/intel_hotplug.c |    3 +-
 5 files changed, 90 insertions(+), 22 deletions(-)
Signature · 6 lines
1.7.10.4

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

PATCH 1/6 drm/i915: add attached connector to hdmi container

From: Shashank Sharma <[email protected]>

This patch adds the intel_connector initialized to intel_hdmi display, during the init phase, just like the other encoders do. This attachment is very useful when we need to extract the connector pointer during the hotplug handler function

---
 drivers/gpu/drm/i915/intel_drv.h  |    1 +
 drivers/gpu/drm/i915/intel_hdmi.c |    1 +
 2 files changed, 2 insertions(+)
drivers/gpu/drm/i915/intel_drv.h +1 -0
index c61ba47..b6c2c20 100644
@@ -667,6 +667,7 @@ struct intel_hdmi {
 	enum hdmi_force_audio force_audio;
 	bool rgb_quant_range_selectable;
 	enum hdmi_picture_aspect aspect_ratio;
+	struct intel_connector *attached_connector;
 	void (*write_infoframe)(struct drm_encoder *encoder,
 				enum hdmi_infoframe_type type,
 				const void *frame, ssize_t len);
drivers/gpu/drm/i915/intel_hdmi.c +1 -0
index ed25f64..5bdb386 100644
@@ -2107,6 +2107,7 @@ void intel_hdmi_init_connector(struct intel_digital_port *intel_dig_port,
 
 	intel_connector_attach_encoder(intel_connector, intel_encoder);
 	drm_connector_register(connector);
+	intel_hdmi->attached_connector = intel_connector;
 
 	/* For G4X desktop chip, PEG_BAND_GAP_DATA 3:0 must first be written
 	 * 0xd.  Failure to do so will result in spurious interrupts being
Signature · 6 lines
1.7.10.4

_______________________________________________
Intel-gfx mailing list
[email protected]
http://lists.freedesktop.org/mailman/listinfo/intel-gfx
Reviewed-by on v1 1/6
#2 lore raw

Re: [PATCH 1/6] drm/i915: add attached connector to hdmi container

Quoted 44 lines · Sonika Jindal

From: Shashank Sharma <[email protected]>

This patch adds the intel_connector initialized to intel_hdmi display, during the init phase, just like the other encoders do. This attachment is very useful when we need to extract the connector pointer during the hotplug handler function

---
 drivers/gpu/drm/i915/intel_drv.h  |    1 +
 drivers/gpu/drm/i915/intel_hdmi.c |    1 +
 2 files changed, 2 insertions(+)
drivers/gpu/drm/i915/intel_drv.h +1 -0
index c61ba47..b6c2c20 100644
@@ -667,6 +667,7 @@ struct intel_hdmi {
        enum hdmi_force_audio force_audio;
        bool rgb_quant_range_selectable;
        enum hdmi_picture_aspect aspect_ratio;
+       struct intel_connector *attached_connector;
        void (*write_infoframe)(struct drm_encoder *encoder,
                                enum hdmi_infoframe_type type,
                                const void *frame, ssize_t len);
drivers/gpu/drm/i915/intel_hdmi.c +1 -0
index ed25f64..5bdb386 100644
@@ -2107,6 +2107,7 @@ void intel_hdmi_init_connector(struct
 intel_digital_port *intel_dig_port,
 
        intel_connector_attach_encoder(intel_connector, intel_encoder);
        drm_connector_register(connector);
+       intel_hdmi->attached_connector = intel_connector;
 
        /* For G4X desktop chip, PEG_BAND_GAP_DATA 3:0 must first be

written
* 0xd. Failure to do so will result in spurious interrupts being
--
1.7.10.4

Quoted 5 lines

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

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

#3 lore raw

Re: [PATCH 1/6] drm/i915: add attached connector to hdmi container

Quoted 13 lines · Rodrigo Vivi
Quoted 8 lines · Sonika Jindal

From: Shashank Sharma <[email protected]>

This patch adds the intel_connector initialized to intel_hdmi display, during the init phase, just like the other encoders do. This attachment is very useful when we need to extract the connector pointer during the hotplug handler function

Queued for -next, thanks for the patch.
-Daniel

Quoted 46 lines
Quoted 36 lines
---
 drivers/gpu/drm/i915/intel_drv.h  |    1 +
 drivers/gpu/drm/i915/intel_hdmi.c |    1 +
 2 files changed, 2 insertions(+)
drivers/gpu/drm/i915/intel_drv.h +1 -0
index c61ba47..b6c2c20 100644
@@ -667,6 +667,7 @@ struct intel_hdmi {
        enum hdmi_force_audio force_audio;
        bool rgb_quant_range_selectable;
        enum hdmi_picture_aspect aspect_ratio;
+       struct intel_connector *attached_connector;
        void (*write_infoframe)(struct drm_encoder *encoder,
                                enum hdmi_infoframe_type type,
                                const void *frame, ssize_t len);
drivers/gpu/drm/i915/intel_hdmi.c +1 -0
index ed25f64..5bdb386 100644
@@ -2107,6 +2107,7 @@ void intel_hdmi_init_connector(struct
 intel_digital_port *intel_dig_port,
 
        intel_connector_attach_encoder(intel_connector, intel_encoder);
        drm_connector_register(connector);
+       intel_hdmi->attached_connector = intel_connector;
 
        /* For G4X desktop chip, PEG_BAND_GAP_DATA 3:0 must first be

written
* 0xd. Failure to do so will result in spurious interrupts being
--
1.7.10.4

Quoted 5 lines

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

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

Signature · 7 lines
Daniel Vetter
Software Engineer, Intel Corporation
http://blog.ffwll.ch
_______________________________________________
Intel-gfx mailing list
[email protected]
http://lists.freedesktop.org/mailman/listinfo/intel-gfx
#4 lore raw

PATCH 2/6 drm/i915: Add HDMI probe function

In reply to Sonika Jindal

From: Shashank Sharma <[email protected]>

This patch adds a separate probe function for HDMI EDID read over DDC channel. This function has been registered as a .hot_plug handler for HDMI encoder.

The current implementation of hdmi_detect()
function re-sets the cached HDMI edid (in connector->detect_edid) in every detect call.This function gets called many times, sometimes directly from userspace probes, forcing drivers to read EDID every detect function call.This causes several problems like:

1. Race conditions in multiple hot_plug / unplug cases, between interrupts bottom halves and userspace detections.
2. Many Un-necessary EDID reads for single hotplug/unplug
3. HDMI complaince failures which expects only one EDID read per hotplug

This function will be serving the purpose of really reading the EDID by really probing the DDC channel, and updating the cached EDID.

The plan is to:
1. i915 IRQ handler bottom half function already calls intel_encoder->hotplug() function. Adding This probe function which will read the EDID only in case of a hotplug / unplug.
2. During init_connector his probe will be called to read the edid
3. Reuse the cached EDID in hdmi_detect() function.

The "< gen7" check is there because this was tested only for >=gen7 platforms. For older platforms the hotplug/reading edid path remains same.

v2: Calling set_edid instead of hdmi_probe during init.
Also, for platforms having DDI, intel_encoder for DP and HDMI is same (taken from intel_dig_port), so for DP also, hot_plug function gets called which is not intended here. So, check for HDMI in intel_hdmi_probe Rely on HPD for updating edid only for platforms gen > 8 and also for VLV.

v3: Dropping the gen < 8 || !VLV check. Now all platforms should rely on hotplug or init for updating the edid.(Daniel)
Also, calling hdmi_probe in init instead of set_edid

---
 drivers/gpu/drm/i915/intel_hdmi.c |   46 ++++++++++++++++++++++++++++++++-----
 1 file changed, 40 insertions(+), 6 deletions(-)
drivers/gpu/drm/i915/intel_hdmi.c +40 -6
index 5bdb386..1eda71a 100644
@@ -1368,23 +1368,53 @@ intel_hdmi_set_edid(struct drm_connector *connector)
 	return connected;
 }
 
+void intel_hdmi_probe(struct intel_encoder *intel_encoder)
+{
+	struct intel_hdmi *intel_hdmi =
+			enc_to_intel_hdmi(&intel_encoder->base);
+	struct intel_connector *intel_connector =
+				intel_hdmi->attached_connector;
+
+	/*
+	 * We are here, means there is a hotplug or a force
+	 * detection. Clear the cached EDID and probe the
+	 * DDC bus to check the current status of HDMI.
+	 */
+	intel_hdmi_unset_edid(&intel_connector->base);
+	if (intel_hdmi_set_edid(&intel_connector->base))
+		DRM_DEBUG_DRIVER("DDC probe: got EDID\n");
+	else
+		DRM_DEBUG_DRIVER("DDC probe: no EDID\n");
+}
+
 static enum drm_connector_status
 intel_hdmi_detect(struct drm_connector *connector, bool force)
 {
 	enum drm_connector_status status;
+	struct intel_connector *intel_connector =
+				to_intel_connector(connector);
 
 	DRM_DEBUG_KMS("[CONNECTOR:%d:%s]\n",
 		      connector->base.id, connector->name);
+	/*
+	 * There are many userspace calls which probe EDID from
+	 * detect path. In case of multiple hotplug/unplug, these
+	 * can cause race conditions while probing EDID. Also its
+	 * waste of CPU cycles to read the EDID again and again
+	 * unless there is a real hotplug.
+	 * So, rely on hotplugs and init to read edid.
+	 * Check connector status based on availability of cached EDID.
+	 */
 
-	intel_hdmi_unset_edid(connector);
-
-	if (intel_hdmi_set_edid(connector)) {
+	if (intel_connector->detect_edid) {
 		struct intel_hdmi *intel_hdmi = intel_attached_hdmi(connector);
-
-		hdmi_to_dig_port(intel_hdmi)->base.type = INTEL_OUTPUT_HDMI;
 		status = connector_status_connected;
-	} else
+		hdmi_to_dig_port(intel_hdmi)->base.type = INTEL_OUTPUT_HDMI;
+		DRM_DEBUG_DRIVER("hdmi status = connected\n");
+	} else {
 		status = connector_status_disconnected;
+		DRM_DEBUG_DRIVER("hdmi status = disconnected\n");
+	}
 
 	return status;
 }
@@ -2104,11 +2134,15 @@ void intel_hdmi_init_connector(struct intel_digital_port *intel_dig_port,
 	intel_connector->unregister = intel_connector_unregister;
 
 	intel_hdmi_add_properties(intel_hdmi, connector);
+	intel_encoder->hot_plug = intel_hdmi_probe;
 
 	intel_connector_attach_encoder(intel_connector, intel_encoder);
 	drm_connector_register(connector);
 	intel_hdmi->attached_connector = intel_connector;
 
+	/* Set edid during init */
+	intel_hdmi_probe(intel_encoder);
+
 	/* For G4X desktop chip, PEG_BAND_GAP_DATA 3:0 must first be written
 	 * 0xd.  Failure to do so will result in spurious interrupts being
 	 * generated on the port when a cable is not attached.
Signature · 6 lines
1.7.10.4

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

Re: [PATCH 2/6] drm/i915: Add HDMI probe function

Quoted 51 lines · Sonika Jindal

From: Shashank Sharma <[email protected]>

This patch adds a separate probe function for HDMI EDID read over DDC channel. This function has been registered as a .hot_plug handler for HDMI encoder.

The current implementation of hdmi_detect()
function re-sets the cached HDMI edid (in connector->detect_edid) in every detect call.This function gets called many times, sometimes directly from userspace probes, forcing drivers to read EDID every detect function call.This causes several problems like:

1. Race conditions in multiple hot_plug / unplug cases, between interrupts bottom halves and userspace detections.
2. Many Un-necessary EDID reads for single hotplug/unplug
3. HDMI complaince failures which expects only one EDID read per hotplug

This function will be serving the purpose of really reading the EDID by really probing the DDC channel, and updating the cached EDID.

The plan is to:
1. i915 IRQ handler bottom half function already calls intel_encoder->hotplug() function. Adding This probe function which will read the EDID only in case of a hotplug / unplug.
2. During init_connector his probe will be called to read the edid
3. Reuse the cached EDID in hdmi_detect() function.

The "< gen7" check is there because this was tested only for >=gen7 platforms. For older platforms the hotplug/reading edid path remains same.

v2: Calling set_edid instead of hdmi_probe during init.
Also, for platforms having DDI, intel_encoder for DP and HDMI is same (taken from intel_dig_port), so for DP also, hot_plug function gets called which is not intended here. So, check for HDMI in intel_hdmi_probe Rely on HPD for updating edid only for platforms gen > 8 and also for VLV.

v3: Dropping the gen < 8 || !VLV check. Now all platforms should rely on hotplug or init for updating the edid.(Daniel)
Also, calling hdmi_probe in init instead of set_edid

---
 drivers/gpu/drm/i915/intel_hdmi.c |   46 ++++++++++++++++++++++++++++++++-----
 1 file changed, 40 insertions(+), 6 deletions(-)
drivers/gpu/drm/i915/intel_hdmi.c +0 -0
index 5bdb386..1eda71a 100644

drivers/gpu/drm/i915/intel_hdmi.c intel_hdmi_set_edid(struct drm_connector *connector) · line 1371

Quoted 4 lines
drivers/gpu/drm/i915/intel_hdmi.c +1 -0
@@ -1368,23 +1368,53 @@ intel_hdmi_set_edid(struct drm_connector *connector)
 	return connected;
 }
 
+void intel_hdmi_probe(struct intel_encoder *intel_encoder)

Please call it _hot_plug if it's for the _hot_plug path.

Quoted 57 lines
drivers/gpu/drm/i915/intel_hdmi.c +35 -6
+{
+	struct intel_hdmi *intel_hdmi =
+			enc_to_intel_hdmi(&intel_encoder->base);
+	struct intel_connector *intel_connector =
+				intel_hdmi->attached_connector;
+
+	/*
+	 * We are here, means there is a hotplug or a force
+	 * detection. Clear the cached EDID and probe the
+	 * DDC bus to check the current status of HDMI.
+	 */
+	intel_hdmi_unset_edid(&intel_connector->base);
+	if (intel_hdmi_set_edid(&intel_connector->base))
+		DRM_DEBUG_DRIVER("DDC probe: got EDID\n");
+	else
+		DRM_DEBUG_DRIVER("DDC probe: no EDID\n");
+}
+
 static enum drm_connector_status
 intel_hdmi_detect(struct drm_connector *connector, bool force)
 {
 	enum drm_connector_status status;
+	struct intel_connector *intel_connector =
+				to_intel_connector(connector);
 
 	DRM_DEBUG_KMS("[CONNECTOR:%d:%s]\n",
 		      connector->base.id, connector->name);
+	/*
+	 * There are many userspace calls which probe EDID from
+	 * detect path. In case of multiple hotplug/unplug, these
+	 * can cause race conditions while probing EDID. Also its
+	 * waste of CPU cycles to read the EDID again and again
+	 * unless there is a real hotplug.
+	 * So, rely on hotplugs and init to read edid.
+	 * Check connector status based on availability of cached EDID.
+	 */
 
-	intel_hdmi_unset_edid(connector);
-
-	if (intel_hdmi_set_edid(connector)) {
+	if (intel_connector->detect_edid) {
 		struct intel_hdmi *intel_hdmi = intel_attached_hdmi(connector);
-
-		hdmi_to_dig_port(intel_hdmi)->base.type = INTEL_OUTPUT_HDMI;
 		status = connector_status_connected;
-	} else
+		hdmi_to_dig_port(intel_hdmi)->base.type = INTEL_OUTPUT_HDMI;
+		DRM_DEBUG_DRIVER("hdmi status = connected\n");
+	} else {
 		status = connector_status_disconnected;
+		DRM_DEBUG_DRIVER("hdmi status = disconnected\n");
+	}
 
 	return status;
 }
@@ -2104,11 +2134,15 @@ void intel_hdmi_init_connector(struct intel_digital_port *intel_dig_port,
 	intel_connector->unregister = intel_connector_unregister;

drivers/gpu/drm/i915/intel_hdmi.c void intel_hdmi_init_connector(struct intel_digital_port *intel_dig_port, · line 2144

Quoted 10 lines
drivers/gpu/drm/i915/intel_hdmi.c +3 -0
 
 	intel_hdmi_add_properties(intel_hdmi, connector);
+	intel_encoder->hot_plug = intel_hdmi_probe;
 
 	intel_connector_attach_encoder(intel_connector, intel_encoder);
 	drm_connector_register(connector);
 	intel_hdmi->attached_connector = intel_connector;
 
+	/* Set edid during init */
+	intel_hdmi_probe(intel_encoder);

Separate patch, but we really don't want this, initial probe should be done async to avoid stalling the driver.
-Daniel

Quoted 11 lines
drivers/gpu/drm/i915/intel_hdmi.c +1 -0
+
 	/* For G4X desktop chip, PEG_BAND_GAP_DATA 3:0 must first be written
 	 * 0xd.  Failure to do so will result in spurious interrupts being
 	 * generated on the port when a cable is not attached.
Signature · 6 lines
1.7.10.4

_______________________________________________
Intel-gfx mailing list
[email protected]
http://lists.freedesktop.org/mailman/listinfo/intel-gfx
Signature · 7 lines
Daniel Vetter
Software Engineer, Intel Corporation
http://blog.ffwll.ch
_______________________________________________
Intel-gfx mailing list
[email protected]
http://lists.freedesktop.org/mailman/listinfo/intel-gfx
Reviewed-by
#6 lore raw

Re: [PATCH 2/6] drm/i915: Add HDMI probe function

I liked the approach and agree with Daniel, so with his suggestions feel free to use:

Quoted 154 lines · Daniel Vetter
Quoted 27 lines · Sonika Jindal

From: Shashank Sharma <[email protected]>

This patch adds a separate probe function for HDMI EDID read over DDC channel. This function has been registered as a .hot_plug handler for HDMI encoder.

The current implementation of hdmi_detect()
function re-sets the cached HDMI edid (in connector->detect_edid) in every detect call.This function gets called many times, sometimes directly from userspace probes, forcing drivers to read EDID every detect function call.This causes several problems like:

1. Race conditions in multiple hot_plug / unplug cases, between interrupts bottom halves and userspace detections.
2. Many Un-necessary EDID reads for single hotplug/unplug
3. HDMI complaince failures which expects only one EDID read per hotplug

This function will be serving the purpose of really reading the EDID by really probing the DDC channel, and updating the cached EDID.

The plan is to:
1. i915 IRQ handler bottom half function already calls intel_encoder->hotplug() function. Adding This probe function which will read the EDID only in case of a hotplug / unplug.
2. During init_connector his probe will be called to read the edid
3. Reuse the cached EDID in hdmi_detect() function.

Quoted 2 lines

The "< gen7" check is there because this was tested only for >=gen7 platforms. For older platforms the hotplug/reading edid path remains

same.

Quoted 4 lines

v2: Calling set_edid instead of hdmi_probe during init. Also, for platforms having DDI, intel_encoder for DP and HDMI is same (taken from intel_dig_port), so for DP also, hot_plug function gets

called

Quoted 2 lines

which is not intended here. So, check for HDMI in intel_hdmi_probe Rely on HPD for updating edid only for platforms gen > 8 and also for

VLV.

Quoted 9 lines

v3: Dropping the gen < 8 || !VLV check. Now all platforms should rely on hotplug or init for updating the edid.(Daniel)
Also, calling hdmi_probe in init instead of set_edid

---
 drivers/gpu/drm/i915/intel_hdmi.c |   46

++++++++++++++++++++++++++++++++-----

Review

Quoted 3 lines
 1 file changed, 40 insertions(+), 6 deletions(-)
(quoted diff) +0 -0

b/drivers/gpu/drm/i915/intel_hdmi.c

drivers/gpu/drm/i915/intel_hdmi.c intel_hdmi_set_edid(struct drm_connector

Quoted 4 lines

index 5bdb386..1eda71a 100644

drivers/gpu/drm/i915/intel_hdmi.c +0 -0
@@ -1368,23 +1368,53 @@ intel_hdmi_set_edid(struct drm_connector

*connector)

drivers/gpu/drm/i915/intel_hdmi.c intel_hdmi_set_edid(struct drm_connector *connector) · line 1371

Quoted 4 lines
drivers/gpu/drm/i915/intel_hdmi.c +1 -0
      return connected;
 }
 
+void intel_hdmi_probe(struct intel_encoder *intel_encoder)

Please call it _hot_plug if it's for the _hot_plug path.

Quoted 32 lines
drivers/gpu/drm/i915/intel_hdmi.c +25 -0
+{
+     struct intel_hdmi *intel_hdmi =
+                     enc_to_intel_hdmi(&intel_encoder->base);
+     struct intel_connector *intel_connector =
+                             intel_hdmi->attached_connector;
+
+     /*
+      * We are here, means there is a hotplug or a force
+      * detection. Clear the cached EDID and probe the
+      * DDC bus to check the current status of HDMI.
+      */
+     intel_hdmi_unset_edid(&intel_connector->base);
+     if (intel_hdmi_set_edid(&intel_connector->base))
+             DRM_DEBUG_DRIVER("DDC probe: got EDID\n");
+     else
+             DRM_DEBUG_DRIVER("DDC probe: no EDID\n");
+}
+
 static enum drm_connector_status
 intel_hdmi_detect(struct drm_connector *connector, bool force)
 {
      enum drm_connector_status status;
+     struct intel_connector *intel_connector =
+                             to_intel_connector(connector);
 
      DRM_DEBUG_KMS("[CONNECTOR:%d:%s]\n",
                    connector->base.id, connector->name);
+     /*
+      * There are many userspace calls which probe EDID from
+      * detect path. In case of multiple hotplug/unplug, these
+      * can cause race conditions while probing EDID. Also its
+      * waste of CPU cycles to read the EDID again and again

drivers/gpu/drm/i915/intel_hdmi.c intel_hdmi_set_edid(struct drm_connector *connector)

Quoted 10 lines
drivers/gpu/drm/i915/intel_hdmi.c +5 -3
+      * unless there is a real hotplug.
+      * So, rely on hotplugs and init to read edid.
+      * Check connector status based on availability of cached EDID.
+      */
 
-     intel_hdmi_unset_edid(connector);
-
-     if (intel_hdmi_set_edid(connector)) {
+     if (intel_connector->detect_edid) {
              struct intel_hdmi *intel_hdmi =

intel_attached_hdmi(connector);

drivers/gpu/drm/i915/intel_hdmi.c intel_hdmi_set_edid(struct drm_connector · line 1384

Quoted 2 lines
drivers/gpu/drm/i915/intel_hdmi.c +0 -2
-
-             hdmi_to_dig_port(intel_hdmi)->base.type =

INTEL_OUTPUT_HDMI;

drivers/gpu/drm/i915/intel_hdmi.c intel_hdmi_set_edid(struct drm_connector *connector)

Quoted 3 lines
drivers/gpu/drm/i915/intel_hdmi.c +1 -1
              status = connector_status_connected;
-     } else
+             hdmi_to_dig_port(intel_hdmi)->base.type =

INTEL_OUTPUT_HDMI;

drivers/gpu/drm/i915/intel_hdmi.c void intel_hdmi_init_connector(struct

Quoted 9 lines
drivers/gpu/drm/i915/intel_hdmi.c +4 -0
+             DRM_DEBUG_DRIVER("hdmi status = connected\n");
+     } else {
              status = connector_status_disconnected;
+             DRM_DEBUG_DRIVER("hdmi status = disconnected\n");
+     }
 
      return status;
 }
@@ -2104,11 +2134,15 @@ void intel_hdmi_init_connector(struct

intel_digital_port *intel_dig_port,

drivers/gpu/drm/i915/intel_hdmi.c void intel_hdmi_init_connector(struct intel_digital_port *intel_dig_port, · line 2144

Quoted 11 lines
drivers/gpu/drm/i915/intel_hdmi.c +3 -0
      intel_connector->unregister = intel_connector_unregister;
 
      intel_hdmi_add_properties(intel_hdmi, connector);
+     intel_encoder->hot_plug = intel_hdmi_probe;
 
      intel_connector_attach_encoder(intel_connector, intel_encoder);
      drm_connector_register(connector);
      intel_hdmi->attached_connector = intel_connector;
 
+     /* Set edid during init */
+     intel_hdmi_probe(intel_encoder);

Separate patch, but we really don't want this, initial probe should be done async to avoid stalling the driver.
-Daniel

drivers/gpu/drm/i915/intel_hdmi.c void intel_hdmi_init_connector(struct · line 2146

Quoted 2 lines
drivers/gpu/drm/i915/intel_hdmi.c +1 -0
+
      /* For G4X desktop chip, PEG_BAND_GAP_DATA 3:0 must first be

written

Quoted 9 lines
drivers/gpu/drm/i915/intel_hdmi.c +0 -1
       * 0xd.  Failure to do so will result in spurious interrupts being
       * generated on the port when a cable is not attached.
--

1.7.10.4

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

Quoted 9 lines

--
Daniel Vetter
Software Engineer, Intel Corporation
http://blog.ffwll.ch
_______________________________________________
Intel-gfx mailing list
[email protected]
http://lists.freedesktop.org/mailman/listinfo/intel-gfx

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

#7 lore raw

Re: [PATCH 2/6] drm/i915: Add HDMI probe function

Quoted 77 lines · Rodrigo Vivi

I liked the approach and agree with Daniel, so with his suggestions feel free to use:

<mailto:[email protected]>>

On Fri, Sep 4, 2015 at 7:46 AM Daniel Vetter <[email protected] <mailto:[email protected]>> wrote:

On Fri, Sep 04, 2015 at 06:56:12PM +0530, Sonika Jindal wrote:
> From: Shashank Sharma <[email protected]
<mailto:[email protected]>>
>
> This patch adds a separate probe function for HDMI
> EDID read over DDC channel. This function has been
> registered as a .hot_plug handler for HDMI encoder.
>
> The current implementation of hdmi_detect()
> function re-sets the cached HDMI edid (in connector->detect_edid) in > every detect call.This function gets called many times, sometimes > directly from userspace probes, forcing drivers to read EDID every > detect function call.This causes several problems like:
>
> 1. Race conditions in multiple hot_plug / unplug cases, between
> interrupts bottom halves and userspace detections.
> 2. Many Un-necessary EDID reads for single hotplug/unplug > 3. HDMI complaince failures which expects only one EDID read per
hotplug
>
> This function will be serving the purpose of really reading the EDID > by really probing the DDC channel, and updating the cached EDID. >
> The plan is to:
> 1. i915 IRQ handler bottom half function already calls
> intel_encoder->hotplug() function. Adding This probe function
which
> will read the EDID only in case of a hotplug / unplug.
> 2. During init_connector his probe will be called to read the edid > 3. Reuse the cached EDID in hdmi_detect() function.
>
> The "< gen7" check is there because this was tested only for >=gen7 > platforms. For older platforms the hotplug/reading edid path
remains same.
>
> v2: Calling set_edid instead of hdmi_probe during init.
> Also, for platforms having DDI, intel_encoder for DP and HDMI is same > (taken from intel_dig_port), so for DP also, hot_plug function
gets called
> which is not intended here. So, check for HDMI in intel_hdmi_probe > Rely on HPD for updating edid only for platforms gen > 8 and also
for VLV.
>
> v3: Dropping the gen < 8 || !VLV check. Now all platforms should
rely on
> hotplug or init for updating the edid.(Daniel)
> Also, calling hdmi_probe in init instead of set_edid
>
> Signed-off-by: Shashank Sharma <[email protected]
<mailto:[email protected]>>
> Signed-off-by: Sonika Jindal <[email protected]
<mailto:[email protected]>>
> ---
> drivers/gpu/drm/i915/intel_hdmi.c | 46
++++++++++++++++++++++++++++++++-----
> 1 file changed, 40 insertions(+), 6 deletions(-)
>
> diff --git a/drivers/gpu/drm/i915/intel_hdmi.c
b/drivers/gpu/drm/i915/intel_hdmi.c
> index 5bdb386..1eda71a 100644
> --- a/drivers/gpu/drm/i915/intel_hdmi.c
> +++ b/drivers/gpu/drm/i915/intel_hdmi.c
> @@ -1368,23 +1368,53 @@ intel_hdmi_set_edid(struct drm_connector
*connector)
> return connected;
> }
>
> +void intel_hdmi_probe(struct intel_encoder *intel_encoder)

Quoted 2 lines

Please call it _hot_plug if it's for the _hot_plug path.

Hmm, initial plan was to call it from other places if any.. But now it seems more logical to name it _hot_plug. I will do that.

Quoted 68 lines
     > +{
     > +     struct intel_hdmi *intel_hdmi =
     > +                     enc_to_intel_hdmi(&intel_encoder->base);
     > +     struct intel_connector *intel_connector =
     > +                             intel_hdmi->attached_connector;
     > +
     > +     /*
     > +      * We are here, means there is a hotplug or a force
     > +      * detection. Clear the cached EDID and probe the
     > +      * DDC bus to check the current status of HDMI.
     > +      */
     > +     intel_hdmi_unset_edid(&intel_connector->base);
     > +     if (intel_hdmi_set_edid(&intel_connector->base))
     > +             DRM_DEBUG_DRIVER("DDC probe: got EDID\n");
     > +     else
     > +             DRM_DEBUG_DRIVER("DDC probe: no EDID\n");
     > +}
     > +
     >  static enum drm_connector_status
     >  intel_hdmi_detect(struct drm_connector *connector, bool force)
     >  {
     >       enum drm_connector_status status;
     > +     struct intel_connector *intel_connector =
     > +                             to_intel_connector(connector);
     >
     >       DRM_DEBUG_KMS("[CONNECTOR:%d:%s]\n",
     >                     connector->base.id <http://base.id>,
    connector->name);
     > +     /*
     > +      * There are many userspace calls which probe EDID from
     > +      * detect path. In case of multiple hotplug/unplug, these
     > +      * can cause race conditions while probing EDID. Also its
     > +      * waste of CPU cycles to read the EDID again and again
     > +      * unless there is a real hotplug.
     > +      * So, rely on hotplugs and init to read edid.
     > +      * Check connector status based on availability of cached EDID.
     > +      */
     >
     > -     intel_hdmi_unset_edid(connector);
     > -
     > -     if (intel_hdmi_set_edid(connector)) {
     > +     if (intel_connector->detect_edid) {
     >               struct intel_hdmi *intel_hdmi =
    intel_attached_hdmi(connector);
     > -
     > -             hdmi_to_dig_port(intel_hdmi)->base.type =
    INTEL_OUTPUT_HDMI;
     >               status = connector_status_connected;
     > -     } else
     > +             hdmi_to_dig_port(intel_hdmi)->base.type =
    INTEL_OUTPUT_HDMI;
     > +             DRM_DEBUG_DRIVER("hdmi status = connected\n");
     > +     } else {
     >               status = connector_status_disconnected;
     > +             DRM_DEBUG_DRIVER("hdmi status = disconnected\n");
     > +     }
     >
     >       return status;
     >  }
     > @@ -2104,11 +2134,15 @@ void intel_hdmi_init_connector(struct
    intel_digital_port *intel_dig_port,
     >       intel_connector->unregister = intel_connector_unregister;
     >
     >       intel_hdmi_add_properties(intel_hdmi, connector);
     > +     intel_encoder->hot_plug = intel_hdmi_probe;
     >
     >       intel_connector_attach_encoder(intel_connector, intel_encoder);
     >       drm_connector_register(connector);
Quoted 8 lines
     >       intel_hdmi->attached_connector = intel_connector;
     >
     > +     /* Set edid during init */
     > +     intel_hdmi_probe(intel_encoder);

    Separate patch, but we really don't want this, initial probe should be
    done async to avoid stalling the driver.
    -Daniel

Where else do you suggest to do this? detect call is one place which does this for DP and used to do for HDMI also. But since we need to take the edid reading part out of edid, I dont see any other place.
Please suggest.

Thanks,
Sonika

Quoted 16 lines
     > +
     >       /* For G4X desktop chip, PEG_BAND_GAP_DATA 3:0 must first
    be written
     >        * 0xd.  Failure to do so will result in spurious
    interrupts being
     >        * generated on the port when a cable is not attached.
     > --
     > 1.7.10.4
     >
     > _______________________________________________
     > Intel-gfx mailing list
     > [email protected]
    <mailto:[email protected]>
     > http://lists.freedesktop.org/mailman/listinfo/intel-gfx
Quoted 9 lines
    --
    Daniel Vetter
    Software Engineer, Intel Corporation
    http://blog.ffwll.ch
    _______________________________________________
    Intel-gfx mailing list
    [email protected] <mailto:[email protected]>
    http://lists.freedesktop.org/mailman/listinfo/intel-gfx

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

#8 lore raw

Re: [PATCH 2/6] drm/i915: Add HDMI probe function

Quoted 163 lines · Sonika
Quoted 72 lines · Rodrigo Vivi

I liked the approach and agree with Daniel, so with his suggestions feel free to use:

<mailto:[email protected]>>

On Fri, Sep 4, 2015 at 7:46 AM Daniel Vetter <[email protected] <mailto:[email protected]>> wrote:

Quoted 1 line · Sonika Jindal

From: Shashank Sharma <[email protected]

<mailto:[email protected]>>

Quoted 11 lines

This patch adds a separate probe function for HDMI EDID read over DDC channel. This function has been registered as a .hot_plug handler for HDMI encoder.

The current implementation of hdmi_detect()
function re-sets the cached HDMI edid (in connector->detect_edid) in every detect call.This function gets called many times, sometimes directly from userspace probes, forcing drivers to read EDID every detect function call.This causes several problems like:

Quoted 4 lines

1. Race conditions in multiple hot_plug / unplug cases, between interrupts bottom halves and userspace detections.
2. Many Un-necessary EDID reads for single hotplug/unplug
3. HDMI complaince failures which expects only one EDID read per

hotplug

Quoted 7 lines

This function will be serving the purpose of really reading the EDID by really probing the DDC channel, and updating the cached EDID.

The plan is to:
1. i915 IRQ handler bottom half function already calls intel_encoder->hotplug() function. Adding This probe function

which

Quoted 6 lines

will read the EDID only in case of a hotplug / unplug.
2. During init_connector his probe will be called to read the edid
3. Reuse the cached EDID in hdmi_detect() function.

The "< gen7" check is there because this was tested only for >=gen7 platforms. For older platforms the hotplug/reading edid path

remains same.

Quoted 4 lines

v2: Calling set_edid instead of hdmi_probe during init. Also, for platforms having DDI, intel_encoder for DP and HDMI is same (taken from intel_dig_port), so for DP also, hot_plug function

gets called

Quoted 2 lines

which is not intended here. So, check for HDMI in intel_hdmi_probe Rely on HPD for updating edid only for platforms gen > 8 and also

for VLV.

Quoted 2 lines

v3: Dropping the gen < 8 || !VLV check. Now all platforms should

rely on

Quoted 4 lines

hotplug or init for updating the edid.(Daniel) Also, calling hdmi_probe in init instead of set_edid

<mailto:[email protected]>>

Quoted 1 line

<mailto:[email protected]>>

Quoted 2 lines
---
 drivers/gpu/drm/i915/intel_hdmi.c |   46

++++++++++++++++++++++++++++++++-----

Review

Quoted 3 lines
 1 file changed, 40 insertions(+), 6 deletions(-)
(quoted diff) +0 -0

b/drivers/gpu/drm/i915/intel_hdmi.c

drivers/gpu/drm/i915/intel_hdmi.c intel_hdmi_set_edid(struct drm_connector

Quoted 4 lines

index 5bdb386..1eda71a 100644

drivers/gpu/drm/i915/intel_hdmi.c +0 -0
@@ -1368,23 +1368,53 @@ intel_hdmi_set_edid(struct drm_connector

*connector)

Quoted 7 lines

drivers/gpu/drm/i915/intel_hdmi.c intel_hdmi_set_edid(struct drm_connector *connector) · line 1371

Quoted 4 lines
drivers/gpu/drm/i915/intel_hdmi.c +1 -0
      return connected;
 }
 
+void intel_hdmi_probe(struct intel_encoder *intel_encoder)

Please call it _hot_plug if it's for the _hot_plug path.

Hmm, initial plan was to call it from other places if any.. But now it seems more logical to name it _hot_plug. I will do that.

Quoted 70 lines
Quoted 17 lines
drivers/gpu/drm/i915/intel_hdmi.c +17 -0
+{
+     struct intel_hdmi *intel_hdmi =
+                     enc_to_intel_hdmi(&intel_encoder->base);
+     struct intel_connector *intel_connector =
+                             intel_hdmi->attached_connector;
+
+     /*
+      * We are here, means there is a hotplug or a force
+      * detection. Clear the cached EDID and probe the
+      * DDC bus to check the current status of HDMI.
+      */
+     intel_hdmi_unset_edid(&intel_connector->base);
+     if (intel_hdmi_set_edid(&intel_connector->base))
+             DRM_DEBUG_DRIVER("DDC probe: got EDID\n");
+     else
+             DRM_DEBUG_DRIVER("DDC probe: no EDID\n");
+}

drivers/gpu/drm/i915/intel_hdmi.c intel_hdmi_set_edid(struct drm_connector *connector)

Quoted 10 lines
drivers/gpu/drm/i915/intel_hdmi.c +3 -0
+
 static enum drm_connector_status
 intel_hdmi_detect(struct drm_connector *connector, bool force)
 {
      enum drm_connector_status status;
+     struct intel_connector *intel_connector =
+                             to_intel_connector(connector);
 
      DRM_DEBUG_KMS("[CONNECTOR:%d:%s]\n",
                    connector->base.id <http://base.id>,

connector->name);

Quoted 5 lines
drivers/gpu/drm/i915/intel_hdmi.c +5 -0
+     /*
+      * There are many userspace calls which probe EDID from
+      * detect path. In case of multiple hotplug/unplug, these
+      * can cause race conditions while probing EDID. Also its
+      * waste of CPU cycles to read the EDID again and again

drivers/gpu/drm/i915/intel_hdmi.c intel_hdmi_set_edid(struct drm_connector *connector)

Quoted 10 lines
drivers/gpu/drm/i915/intel_hdmi.c +5 -3
+      * unless there is a real hotplug.
+      * So, rely on hotplugs and init to read edid.
+      * Check connector status based on availability of cached EDID.
+      */
 
-     intel_hdmi_unset_edid(connector);
-
-     if (intel_hdmi_set_edid(connector)) {
+     if (intel_connector->detect_edid) {
              struct intel_hdmi *intel_hdmi =
   intel_attached_hdmi(connector);
    > -
    > -             hdmi_to_dig_port(intel_hdmi)->base.type =
   INTEL_OUTPUT_HDMI;
    >               status = connector_status_connected;
    > -     } else
    > +             hdmi_to_dig_port(intel_hdmi)->base.type =
   INTEL_OUTPUT_HDMI;
    > +             DRM_DEBUG_DRIVER("hdmi status = connected\n");
    > +     } else {
    >               status = connector_status_disconnected;
    > +             DRM_DEBUG_DRIVER("hdmi status = disconnected\n");
    > +     }
    >
    >       return status;
    >  }
    > @@ -2104,11 +2134,15 @@ void intel_hdmi_init_connector(struct

intel_digital_port *intel_dig_port,

Quoted 9 lines
drivers/gpu/drm/i915/intel_hdmi.c +1 -0
      intel_connector->unregister = intel_connector_unregister;
 
      intel_hdmi_add_properties(intel_hdmi, connector);
+     intel_encoder->hot_plug = intel_hdmi_probe;
 
      intel_connector_attach_encoder(intel_connector, intel_encoder);
      drm_connector_register(connector);
      intel_hdmi->attached_connector = intel_connector;
 
Quoted 6 lines

drivers/gpu/drm/i915/intel_hdmi.c void intel_hdmi_init_connector(struct intel_digital_port *intel_dig_port, · line 2144

Quoted 2 lines
drivers/gpu/drm/i915/intel_hdmi.c +2 -0
+     /* Set edid during init */
+     intel_hdmi_probe(intel_encoder);

Separate patch, but we really don't want this, initial probe should be done async to avoid stalling the driver.
-Daniel

Quoted 5 lines

Where else do you suggest to do this? detect call is one place which does this for DP and used to do for HDMI also.
But since we need to take the edid reading part out of edid, I dont see any other place.
Please suggest.

You need to inject a fake hpd or similar everytime the world might have changed from under us, which isn't just init but also system resume.

We do that already, so you just have to make sure that keeps on working. Probably another reason why it would be better to rework that entire hpd handling/detection code, to make sure it's clean and can be understood.

Adding hacks like these in the init code is certainly the wrong approach, since we never done detection this early (so surprising from a high-level ordering) and it also misses resume (which is why it should be done in code shared between driver load and resume).
-Daniel

Signature · 7 lines
Daniel Vetter
Software Engineer, Intel Corporation
http://blog.ffwll.ch
_______________________________________________
Intel-gfx mailing list
[email protected]
http://lists.freedesktop.org/mailman/listinfo/intel-gfx
#9 lore raw

Re: [PATCH 2/6] drm/i915: Add HDMI probe function

Quoted 1 line

You need to inject a fake hpd or similar everytime the world might have changed from under us, which isn't just init but also system resume.

This is a fake hpd only, just without wrapper, if you want, we can create a wrapper function like (intel_hdmi_insert_fake_hpd), and call this function from init() And call HDMI probe(hot_plug) function from here. This would be just adding some CPU cycles in the code.

Quoted 2 lines

Adding hacks like these in the init code is certainly the wrong approach, since we never done detection this early (so surprising from a high-level ordering) and it also misses resume (which is why it should be done in code shared between driver load and resume).

During the ADF code development, we saw that fb_console comes up very soon, as soon as the intel_fbdev is ready, causing a detect() on a connected HDMI. Do you think intel_hpd_init() would be the right place to probe, coz as you rightly mentioned, we will need this on suspend/resume also.

Regards
Shashank
-----Original Message-----
From: Intel-gfx [mailto:[email protected]] On Behalf Of Daniel Vetter
Sent: Monday, September 14, 2015 2:00 PM
To: Jindal, Sonika
Cc: [email protected]
Subject: Re: [Intel-gfx] [PATCH 2/6] drm/i915: Add HDMI probe function

Quoted 163 lines · Sonika
Quoted 72 lines · Rodrigo Vivi

I liked the approach and agree with Daniel, so with his suggestions feel free to use:

<mailto:[email protected]>>

On Fri, Sep 4, 2015 at 7:46 AM Daniel Vetter <[email protected] <mailto:[email protected]>> wrote:

Quoted 1 line · Sonika Jindal

From: Shashank Sharma <[email protected]

<mailto:[email protected]>>

Quoted 11 lines

This patch adds a separate probe function for HDMI EDID read over DDC channel. This function has been registered as a .hot_plug handler for HDMI encoder.

The current implementation of hdmi_detect()
function re-sets the cached HDMI edid (in connector->detect_edid) in every detect call.This function gets called many times, sometimes directly from userspace probes, forcing drivers to read EDID every detect function call.This causes several problems like:

Quoted 4 lines

1. Race conditions in multiple hot_plug / unplug cases, between interrupts bottom halves and userspace detections.
2. Many Un-necessary EDID reads for single hotplug/unplug
3. HDMI complaince failures which expects only one EDID read per

hotplug

Quoted 7 lines

This function will be serving the purpose of really reading the EDID by really probing the DDC channel, and updating the cached EDID.

The plan is to:
1. i915 IRQ handler bottom half function already calls intel_encoder->hotplug() function. Adding This probe function

which

Quoted 6 lines

will read the EDID only in case of a hotplug / unplug.
2. During init_connector his probe will be called to read the edid
3. Reuse the cached EDID in hdmi_detect() function.

The "< gen7" check is there because this was tested only for >=gen7 platforms. For older platforms the hotplug/reading edid path

remains same.

Quoted 4 lines

v2: Calling set_edid instead of hdmi_probe during init. Also, for platforms having DDI, intel_encoder for DP and HDMI is same (taken from intel_dig_port), so for DP also, hot_plug function

gets called

Quoted 2 lines

which is not intended here. So, check for HDMI in intel_hdmi_probe Rely on HPD for updating edid only for platforms gen > 8 and also

for VLV.

Quoted 2 lines

v3: Dropping the gen < 8 || !VLV check. Now all platforms should

rely on

Quoted 4 lines

hotplug or init for updating the edid.(Daniel) Also, calling hdmi_probe in init instead of set_edid

<mailto:[email protected]>>

Quoted 1 line

<mailto:[email protected]>>

Quoted 2 lines
---
 drivers/gpu/drm/i915/intel_hdmi.c |   46

++++++++++++++++++++++++++++++++-----

Review

Quoted 3 lines
 1 file changed, 40 insertions(+), 6 deletions(-)
(quoted diff) +0 -0

b/drivers/gpu/drm/i915/intel_hdmi.c

drivers/gpu/drm/i915/intel_hdmi.c intel_hdmi_set_edid(struct drm_connector

Quoted 4 lines

index 5bdb386..1eda71a 100644

drivers/gpu/drm/i915/intel_hdmi.c +0 -0
@@ -1368,23 +1368,53 @@ intel_hdmi_set_edid(struct drm_connector

*connector)

Quoted 7 lines

drivers/gpu/drm/i915/intel_hdmi.c intel_hdmi_set_edid(struct drm_connector *connector) · line 1371

Quoted 4 lines
drivers/gpu/drm/i915/intel_hdmi.c +1 -0
      return connected;
 }
 
+void intel_hdmi_probe(struct intel_encoder *intel_encoder)

Please call it _hot_plug if it's for the _hot_plug path.

Hmm, initial plan was to call it from other places if any.. But now it seems more logical to name it _hot_plug. I will do that.

Quoted 70 lines
Quoted 17 lines
drivers/gpu/drm/i915/intel_hdmi.c +17 -0
+{
+     struct intel_hdmi *intel_hdmi =
+                     enc_to_intel_hdmi(&intel_encoder->base);
+     struct intel_connector *intel_connector =
+                             intel_hdmi->attached_connector;
+
+     /*
+      * We are here, means there is a hotplug or a force
+      * detection. Clear the cached EDID and probe the
+      * DDC bus to check the current status of HDMI.
+      */
+     intel_hdmi_unset_edid(&intel_connector->base);
+     if (intel_hdmi_set_edid(&intel_connector->base))
+             DRM_DEBUG_DRIVER("DDC probe: got EDID\n");
+     else
+             DRM_DEBUG_DRIVER("DDC probe: no EDID\n");
+}

drivers/gpu/drm/i915/intel_hdmi.c intel_hdmi_set_edid(struct drm_connector *connector)

Quoted 10 lines
drivers/gpu/drm/i915/intel_hdmi.c +3 -0
+
 static enum drm_connector_status
 intel_hdmi_detect(struct drm_connector *connector, bool force)
 {
      enum drm_connector_status status;
+     struct intel_connector *intel_connector =
+                             to_intel_connector(connector);
 
      DRM_DEBUG_KMS("[CONNECTOR:%d:%s]\n",
                    connector->base.id <http://base.id>,

connector->name);

Quoted 5 lines
drivers/gpu/drm/i915/intel_hdmi.c +5 -0
+     /*
+      * There are many userspace calls which probe EDID from
+      * detect path. In case of multiple hotplug/unplug, these
+      * can cause race conditions while probing EDID. Also its
+      * waste of CPU cycles to read the EDID again and again

drivers/gpu/drm/i915/intel_hdmi.c intel_hdmi_set_edid(struct drm_connector *connector)

Quoted 10 lines
drivers/gpu/drm/i915/intel_hdmi.c +5 -3
+      * unless there is a real hotplug.
+      * So, rely on hotplugs and init to read edid.
+      * Check connector status based on availability of cached EDID.
+      */
 
-     intel_hdmi_unset_edid(connector);
-
-     if (intel_hdmi_set_edid(connector)) {
+     if (intel_connector->detect_edid) {
              struct intel_hdmi *intel_hdmi =
   intel_attached_hdmi(connector);
    > -
    > -             hdmi_to_dig_port(intel_hdmi)->base.type =
   INTEL_OUTPUT_HDMI;
    >               status = connector_status_connected;
    > -     } else
    > +             hdmi_to_dig_port(intel_hdmi)->base.type =
   INTEL_OUTPUT_HDMI;
    > +             DRM_DEBUG_DRIVER("hdmi status = connected\n");
    > +     } else {
    >               status = connector_status_disconnected;
    > +             DRM_DEBUG_DRIVER("hdmi status = disconnected\n");
    > +     }
    >
    >       return status;
    >  }
    > @@ -2104,11 +2134,15 @@ void intel_hdmi_init_connector(struct

intel_digital_port *intel_dig_port,

Quoted 9 lines
drivers/gpu/drm/i915/intel_hdmi.c +1 -0
      intel_connector->unregister = intel_connector_unregister;
 
      intel_hdmi_add_properties(intel_hdmi, connector);
+     intel_encoder->hot_plug = intel_hdmi_probe;
 
      intel_connector_attach_encoder(intel_connector, intel_encoder);
      drm_connector_register(connector);
      intel_hdmi->attached_connector = intel_connector;
 
Quoted 6 lines

drivers/gpu/drm/i915/intel_hdmi.c void intel_hdmi_init_connector(struct intel_digital_port *intel_dig_port, · line 2144

Quoted 2 lines
drivers/gpu/drm/i915/intel_hdmi.c +2 -0
+     /* Set edid during init */
+     intel_hdmi_probe(intel_encoder);

Separate patch, but we really don't want this, initial probe should be done async to avoid stalling the driver.
-Daniel

Quoted 5 lines

Where else do you suggest to do this? detect call is one place which does this for DP and used to do for HDMI also.
But since we need to take the edid reading part out of edid, I dont see any other place.
Please suggest.

You need to inject a fake hpd or similar everytime the world might have changed from under us, which isn't just init but also system resume.

We do that already, so you just have to make sure that keeps on working.
Probably another reason why it would be better to rework that entire hpd handling/detection code, to make sure it's clean and can be understood.

Adding hacks like these in the init code is certainly the wrong approach, since we never done detection this early (so surprising from a high-level ordering) and it also misses resume (which is why it should be done in code shared between driver load and resume).
-Daniel
--
Daniel Vetter
Software Engineer, Intel Corporation
http://blog.ffwll.ch
_______________________________________________
Intel-gfx mailing list
[email protected]
http://lists.freedesktop.org/mailman/listinfo/intel-gfx
_______________________________________________
Intel-gfx mailing list
[email protected]
http://lists.freedesktop.org/mailman/listinfo/intel-gfx

#10 lore raw

Re: [PATCH 2/6] drm/i915: Add HDMI probe function

Quoted 5 lines · Shashank
Quoted 1 line

You need to inject a fake hpd or similar everytime the world might have changed from under us, which isn't just init but also system resume.

This is a fake hpd only, just without wrapper, if you want, we can create a wrapper function like (intel_hdmi_insert_fake_hpd), and call this function from init() And call HDMI probe(hot_plug) function from here. This would be just adding some CPU cycles in the code.

We already have fake hpd handling, and it's done at a higher level (and in all the correct places) using drm_helper_hpd_irq_event.

Your patch here breaks that, but then only partially fixes things up again (you're missing system resume). That's not a good design, since it means we need bandaids all over also in the future for anything else (like DP).

Hence imo why we should create a new overall design with how this is supposed to be handled, and then change the core bits (so probably we need an intel version of drm_helper_hpd_irq_event and use that instead of the current calls). Eventually we want to make that async to speed up resume.

Quoted 7 lines
Quoted 2 lines

Adding hacks like these in the init code is certainly the wrong approach, since we never done detection this early (so surprising from a high-level ordering) and it also misses resume (which is why it should be done in code shared between driver load and resume).

During the ADF code development, we saw that fb_console comes up very soon, as soon as the intel_fbdev is ready, causing a detect() on a connected HDMI. Do you think intel_hpd_init() would be the right place to probe, coz as you rightly mentioned, we will need this on suspend/resume also.

Nope. fbdev is optional and you can't rely upon that, or write the code assuming it's there. The initial probing at driver load time is atm not done at all, we rely upon fbdev or early userspace to do the probing.

What you probably want to do instead is reset the edid cache on driver load and on resume to make sure we don't cache a potentially bogus edid. We have the encoder->reset hook for that kind of work. That might be another approach.

But doing probe in _init for dynamic ports isnt' the right approach imo. -Daniel

Quoted 196 lines

Regards
Shashank
-----Original Message-----
From: Intel-gfx [mailto:[email protected]] On Behalf Of Daniel Vetter
Sent: Monday, September 14, 2015 2:00 PM
To: Jindal, Sonika
Cc: [email protected]
Subject: Re: [Intel-gfx] [PATCH 2/6] drm/i915: Add HDMI probe function

Quoted 163 lines · Sonika
Quoted 72 lines · Rodrigo Vivi

I liked the approach and agree with Daniel, so with his suggestions feel free to use:

<mailto:[email protected]>>

On Fri, Sep 4, 2015 at 7:46 AM Daniel Vetter <[email protected] <mailto:[email protected]>> wrote:

Quoted 1 line · Sonika Jindal

From: Shashank Sharma <[email protected]

<mailto:[email protected]>>

Quoted 11 lines

This patch adds a separate probe function for HDMI EDID read over DDC channel. This function has been registered as a .hot_plug handler for HDMI encoder.

The current implementation of hdmi_detect()
function re-sets the cached HDMI edid (in connector->detect_edid) in every detect call.This function gets called many times, sometimes directly from userspace probes, forcing drivers to read EDID every detect function call.This causes several problems like:

Quoted 4 lines

1. Race conditions in multiple hot_plug / unplug cases, between interrupts bottom halves and userspace detections.
2. Many Un-necessary EDID reads for single hotplug/unplug
3. HDMI complaince failures which expects only one EDID read per

hotplug

Quoted 7 lines

This function will be serving the purpose of really reading the EDID by really probing the DDC channel, and updating the cached EDID.

The plan is to:
1. i915 IRQ handler bottom half function already calls intel_encoder->hotplug() function. Adding This probe function

which

Quoted 6 lines

will read the EDID only in case of a hotplug / unplug.
2. During init_connector his probe will be called to read the edid
3. Reuse the cached EDID in hdmi_detect() function.

The "< gen7" check is there because this was tested only for >=gen7 platforms. For older platforms the hotplug/reading edid path

remains same.

Quoted 4 lines

v2: Calling set_edid instead of hdmi_probe during init. Also, for platforms having DDI, intel_encoder for DP and HDMI is same (taken from intel_dig_port), so for DP also, hot_plug function

gets called

Quoted 2 lines

which is not intended here. So, check for HDMI in intel_hdmi_probe Rely on HPD for updating edid only for platforms gen > 8 and also

for VLV.

Quoted 2 lines

v3: Dropping the gen < 8 || !VLV check. Now all platforms should

rely on

Quoted 4 lines

hotplug or init for updating the edid.(Daniel) Also, calling hdmi_probe in init instead of set_edid

<mailto:[email protected]>>

Quoted 1 line

<mailto:[email protected]>>

Quoted 2 lines
---
 drivers/gpu/drm/i915/intel_hdmi.c |   46

++++++++++++++++++++++++++++++++-----

Review

Quoted 3 lines
 1 file changed, 40 insertions(+), 6 deletions(-)
(quoted diff) +0 -0

b/drivers/gpu/drm/i915/intel_hdmi.c

drivers/gpu/drm/i915/intel_hdmi.c intel_hdmi_set_edid(struct drm_connector

Quoted 4 lines

index 5bdb386..1eda71a 100644

drivers/gpu/drm/i915/intel_hdmi.c +0 -0
@@ -1368,23 +1368,53 @@ intel_hdmi_set_edid(struct drm_connector

*connector)

Quoted 7 lines

drivers/gpu/drm/i915/intel_hdmi.c intel_hdmi_set_edid(struct drm_connector *connector) · line 1371

Quoted 4 lines
drivers/gpu/drm/i915/intel_hdmi.c +1 -0
      return connected;
 }
 
+void intel_hdmi_probe(struct intel_encoder *intel_encoder)

Please call it _hot_plug if it's for the _hot_plug path.

Hmm, initial plan was to call it from other places if any.. But now it seems more logical to name it _hot_plug. I will do that.

Quoted 70 lines
Quoted 17 lines
drivers/gpu/drm/i915/intel_hdmi.c +17 -0
+{
+     struct intel_hdmi *intel_hdmi =
+                     enc_to_intel_hdmi(&intel_encoder->base);
+     struct intel_connector *intel_connector =
+                             intel_hdmi->attached_connector;
+
+     /*
+      * We are here, means there is a hotplug or a force
+      * detection. Clear the cached EDID and probe the
+      * DDC bus to check the current status of HDMI.
+      */
+     intel_hdmi_unset_edid(&intel_connector->base);
+     if (intel_hdmi_set_edid(&intel_connector->base))
+             DRM_DEBUG_DRIVER("DDC probe: got EDID\n");
+     else
+             DRM_DEBUG_DRIVER("DDC probe: no EDID\n");
+}

drivers/gpu/drm/i915/intel_hdmi.c intel_hdmi_set_edid(struct drm_connector *connector)

Quoted 10 lines
drivers/gpu/drm/i915/intel_hdmi.c +3 -0
+
 static enum drm_connector_status
 intel_hdmi_detect(struct drm_connector *connector, bool force)
 {
      enum drm_connector_status status;
+     struct intel_connector *intel_connector =
+                             to_intel_connector(connector);
 
      DRM_DEBUG_KMS("[CONNECTOR:%d:%s]\n",
                    connector->base.id <http://base.id>,

connector->name);

Quoted 5 lines
drivers/gpu/drm/i915/intel_hdmi.c +5 -0
+     /*
+      * There are many userspace calls which probe EDID from
+      * detect path. In case of multiple hotplug/unplug, these
+      * can cause race conditions while probing EDID. Also its
+      * waste of CPU cycles to read the EDID again and again

drivers/gpu/drm/i915/intel_hdmi.c intel_hdmi_set_edid(struct drm_connector *connector)

Quoted 10 lines
drivers/gpu/drm/i915/intel_hdmi.c +5 -3
+      * unless there is a real hotplug.
+      * So, rely on hotplugs and init to read edid.
+      * Check connector status based on availability of cached EDID.
+      */
 
-     intel_hdmi_unset_edid(connector);
-
-     if (intel_hdmi_set_edid(connector)) {
+     if (intel_connector->detect_edid) {
              struct intel_hdmi *intel_hdmi =
   intel_attached_hdmi(connector);
    > -
    > -             hdmi_to_dig_port(intel_hdmi)->base.type =
   INTEL_OUTPUT_HDMI;
    >               status = connector_status_connected;
    > -     } else
    > +             hdmi_to_dig_port(intel_hdmi)->base.type =
   INTEL_OUTPUT_HDMI;
    > +             DRM_DEBUG_DRIVER("hdmi status = connected\n");
    > +     } else {
    >               status = connector_status_disconnected;
    > +             DRM_DEBUG_DRIVER("hdmi status = disconnected\n");
    > +     }
    >
    >       return status;
    >  }
    > @@ -2104,11 +2134,15 @@ void intel_hdmi_init_connector(struct

intel_digital_port *intel_dig_port,

Quoted 9 lines
drivers/gpu/drm/i915/intel_hdmi.c +1 -0
      intel_connector->unregister = intel_connector_unregister;
 
      intel_hdmi_add_properties(intel_hdmi, connector);
+     intel_encoder->hot_plug = intel_hdmi_probe;
 
      intel_connector_attach_encoder(intel_connector, intel_encoder);
      drm_connector_register(connector);
      intel_hdmi->attached_connector = intel_connector;
 
Quoted 6 lines

drivers/gpu/drm/i915/intel_hdmi.c void intel_hdmi_init_connector(struct intel_digital_port *intel_dig_port, · line 2144

Quoted 2 lines
drivers/gpu/drm/i915/intel_hdmi.c +2 -0
+     /* Set edid during init */
+     intel_hdmi_probe(intel_encoder);

Separate patch, but we really don't want this, initial probe should be done async to avoid stalling the driver.
-Daniel

Quoted 5 lines

Where else do you suggest to do this? detect call is one place which does this for DP and used to do for HDMI also.
But since we need to take the edid reading part out of edid, I dont see any other place.
Please suggest.

You need to inject a fake hpd or similar everytime the world might have changed from under us, which isn't just init but also system resume.

We do that already, so you just have to make sure that keeps on working.
Probably another reason why it would be better to rework that entire hpd handling/detection code, to make sure it's clean and can be understood.

Adding hacks like these in the init code is certainly the wrong approach, since we never done detection this early (so surprising from a high-level ordering) and it also misses resume (which is why it should be done in code shared between driver load and resume).
-Daniel
--
Daniel Vetter
Software Engineer, Intel Corporation
http://blog.ffwll.ch
_______________________________________________
Intel-gfx mailing list
[email protected]
http://lists.freedesktop.org/mailman/listinfo/intel-gfx

Signature · 7 lines
Daniel Vetter
Software Engineer, Intel Corporation
http://blog.ffwll.ch
_______________________________________________
Intel-gfx mailing list
[email protected]
http://lists.freedesktop.org/mailman/listinfo/intel-gfx
#11 lore raw

Re: [PATCH 2/6] drm/i915: Add HDMI probe function

Regards
Shashank

Quoted 10 lines · Daniel Vetter
Quoted 5 lines · Shashank
Quoted 1 line

You need to inject a fake hpd or similar everytime the world might have changed from under us, which isn't just init but also system resume.

This is a fake hpd only, just without wrapper, if you want, we can create a wrapper function like (intel_hdmi_insert_fake_hpd), and call this function from init() And call HDMI probe(hot_plug) function from here. This would be just adding some CPU cycles in the code.

We already have fake hpd handling, and it's done at a higher level (and in all the correct places) using drm_helper_hpd_irq_event.

Thanks, this sounds like a good suggestion here.
We can call the hot_plug() from drm_helper_hpd_irq_event, just before it calls the detect(), and our state machine is still intact.

Quoted 4 lines

Your patch here breaks that, but then only partially fixes things up again (you're missing system resume). That's not a good design, since it means we need bandaids all over also in the future for anything else (like DP).

I guess, in this was we wont break anything existing.

Quoted 13 lines

Hence imo why we should create a new overall design with how this is supposed to be handled, and then change the core bits (so probably we need an intel version of drm_helper_hpd_irq_event and use that instead of the current calls). Eventually we want to make that async to speed up resume.

Quoted 7 lines
Quoted 2 lines

Adding hacks like these in the init code is certainly the wrong approach, since we never done detection this early (so surprising from a high-level ordering) and it also misses resume (which is why it should be done in code shared between driver load and resume).

During the ADF code development, we saw that fb_console comes up very soon, as soon as the intel_fbdev is ready, causing a detect() on a connected HDMI. Do you think intel_hpd_init() would be the right place to probe, coz as you rightly mentioned, we will need this on suspend/resume also.

Quoted 9 lines

Nope. fbdev is optional and you can't rely upon that, or write the code assuming it's there. The initial probing at driver load time is atm not done at all, we rely upon fbdev or early userspace to do the probing.

What you probably want to do instead is reset the edid cache on driver load and on resume to make sure we don't cache a potentially bogus edid. We have the encoder->reset hook for that kind of work. That might be another approach.

Yes, thats exactly what hot_plug function does. It checks the live status and if we have cached EDID, it releases the EDID. Now its just matter of calling this from the right place, in the right way.

Quoted 3 lines

But doing probe in _init for dynamic ports isnt' the right approach imo. -Daniel

Yep, agree. Let us work on this part.

Quoted 185 lines
Quoted 185 lines

Regards
Shashank
-----Original Message-----
From: Intel-gfx [mailto:[email protected]] On Behalf Of Daniel Vetter
Sent: Monday, September 14, 2015 2:00 PM
To: Jindal, Sonika
Cc: [email protected]
Subject: Re: [Intel-gfx] [PATCH 2/6] drm/i915: Add HDMI probe function

Quoted 163 lines · Sonika
Quoted 77 lines · Rodrigo Vivi

I liked the approach and agree with Daniel, so with his suggestions feel free to use:

<mailto:[email protected]>>

On Fri, Sep 4, 2015 at 7:46 AM Daniel Vetter <[email protected] <mailto:[email protected]>> wrote:

On Fri, Sep 04, 2015 at 06:56:12PM +0530, Sonika Jindal wrote:
> From: Shashank Sharma <[email protected]
<mailto:[email protected]>>
>
> This patch adds a separate probe function for HDMI
> EDID read over DDC channel. This function has been
> registered as a .hot_plug handler for HDMI encoder.
>
> The current implementation of hdmi_detect()
> function re-sets the cached HDMI edid (in connector->detect_edid) in > every detect call.This function gets called many times, sometimes > directly from userspace probes, forcing drivers to read EDID every > detect function call.This causes several problems like:
>
> 1. Race conditions in multiple hot_plug / unplug cases, between
> interrupts bottom halves and userspace detections.
> 2. Many Un-necessary EDID reads for single hotplug/unplug > 3. HDMI complaince failures which expects only one EDID read per
hotplug
>
> This function will be serving the purpose of really reading the EDID > by really probing the DDC channel, and updating the cached EDID. >
> The plan is to:
> 1. i915 IRQ handler bottom half function already calls
> intel_encoder->hotplug() function. Adding This probe function
which
> will read the EDID only in case of a hotplug / unplug.
> 2. During init_connector his probe will be called to read the edid > 3. Reuse the cached EDID in hdmi_detect() function.
>
> The "< gen7" check is there because this was tested only for >=gen7 > platforms. For older platforms the hotplug/reading edid path
remains same.
>
> v2: Calling set_edid instead of hdmi_probe during init.
> Also, for platforms having DDI, intel_encoder for DP and HDMI is same > (taken from intel_dig_port), so for DP also, hot_plug function
gets called
> which is not intended here. So, check for HDMI in intel_hdmi_probe > Rely on HPD for updating edid only for platforms gen > 8 and also
for VLV.
>
> v3: Dropping the gen < 8 || !VLV check. Now all platforms should
rely on
> hotplug or init for updating the edid.(Daniel)
> Also, calling hdmi_probe in init instead of set_edid
>
> Signed-off-by: Shashank Sharma <[email protected]
<mailto:[email protected]>>
> Signed-off-by: Sonika Jindal <[email protected]
<mailto:[email protected]>>
> ---
> drivers/gpu/drm/i915/intel_hdmi.c | 46
++++++++++++++++++++++++++++++++-----
> 1 file changed, 40 insertions(+), 6 deletions(-)
>
> diff --git a/drivers/gpu/drm/i915/intel_hdmi.c
b/drivers/gpu/drm/i915/intel_hdmi.c
> index 5bdb386..1eda71a 100644
> --- a/drivers/gpu/drm/i915/intel_hdmi.c
> +++ b/drivers/gpu/drm/i915/intel_hdmi.c
> @@ -1368,23 +1368,53 @@ intel_hdmi_set_edid(struct drm_connector
*connector)
> return connected;
> }
>
> +void intel_hdmi_probe(struct intel_encoder *intel_encoder)

Quoted 2 lines

Please call it _hot_plug if it's for the _hot_plug path.

Hmm, initial plan was to call it from other places if any.. But now it seems more logical to name it _hot_plug. I will do that.

Quoted 68 lines
     > +{
     > +     struct intel_hdmi *intel_hdmi =
     > +                     enc_to_intel_hdmi(&intel_encoder->base);
     > +     struct intel_connector *intel_connector =
     > +                             intel_hdmi->attached_connector;
     > +
     > +     /*
     > +      * We are here, means there is a hotplug or a force
     > +      * detection. Clear the cached EDID and probe the
     > +      * DDC bus to check the current status of HDMI.
     > +      */
     > +     intel_hdmi_unset_edid(&intel_connector->base);
     > +     if (intel_hdmi_set_edid(&intel_connector->base))
     > +             DRM_DEBUG_DRIVER("DDC probe: got EDID\n");
     > +     else
     > +             DRM_DEBUG_DRIVER("DDC probe: no EDID\n");
     > +}
     > +
     >  static enum drm_connector_status
     >  intel_hdmi_detect(struct drm_connector *connector, bool force)
     >  {
     >       enum drm_connector_status status;
     > +     struct intel_connector *intel_connector =
     > +                             to_intel_connector(connector);
     >
     >       DRM_DEBUG_KMS("[CONNECTOR:%d:%s]\n",
     >                     connector->base.id <http://base.id>,
    connector->name);
     > +     /*
     > +      * There are many userspace calls which probe EDID from
     > +      * detect path. In case of multiple hotplug/unplug, these
     > +      * can cause race conditions while probing EDID. Also its
     > +      * waste of CPU cycles to read the EDID again and again
     > +      * unless there is a real hotplug.
     > +      * So, rely on hotplugs and init to read edid.
     > +      * Check connector status based on availability of cached EDID.
     > +      */
     >
     > -     intel_hdmi_unset_edid(connector);
     > -
     > -     if (intel_hdmi_set_edid(connector)) {
     > +     if (intel_connector->detect_edid) {
     >               struct intel_hdmi *intel_hdmi =
    intel_attached_hdmi(connector);
     > -
     > -             hdmi_to_dig_port(intel_hdmi)->base.type =
    INTEL_OUTPUT_HDMI;
     >               status = connector_status_connected;
     > -     } else
     > +             hdmi_to_dig_port(intel_hdmi)->base.type =
    INTEL_OUTPUT_HDMI;
     > +             DRM_DEBUG_DRIVER("hdmi status = connected\n");
     > +     } else {
     >               status = connector_status_disconnected;
     > +             DRM_DEBUG_DRIVER("hdmi status = disconnected\n");
     > +     }
     >
     >       return status;
     >  }
     > @@ -2104,11 +2134,15 @@ void intel_hdmi_init_connector(struct
    intel_digital_port *intel_dig_port,
     >       intel_connector->unregister = intel_connector_unregister;
     >
     >       intel_hdmi_add_properties(intel_hdmi, connector);
     > +     intel_encoder->hot_plug = intel_hdmi_probe;
     >
     >       intel_connector_attach_encoder(intel_connector, intel_encoder);
     >       drm_connector_register(connector);
Quoted 8 lines
     >       intel_hdmi->attached_connector = intel_connector;
     >
     > +     /* Set edid during init */
     > +     intel_hdmi_probe(intel_encoder);

    Separate patch, but we really don't want this, initial probe should be
    done async to avoid stalling the driver.
    -Daniel
Quoted 5 lines

Where else do you suggest to do this? detect call is one place which does this for DP and used to do for HDMI also.
But since we need to take the edid reading part out of edid, I dont see any other place.
Please suggest.

You need to inject a fake hpd or similar everytime the world might have changed from under us, which isn't just init but also system resume.

We do that already, so you just have to make sure that keeps on working.
Probably another reason why it would be better to rework that entire hpd handling/detection code, to make sure it's clean and can be understood.

Quoted 12 lines
Quoted 11 lines

Adding hacks like these in the init code is certainly the wrong approach, since we never done detection this early (so surprising from a high-level ordering) and it also misses resume (which is why it should be done in code shared between driver load and resume).
-Daniel
--
Daniel Vetter
Software Engineer, Intel Corporation
http://blog.ffwll.ch
_______________________________________________
Intel-gfx mailing list
[email protected]
http://lists.freedesktop.org/mailman/listinfo/intel-gfx

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

#12 lore raw

PATCH 1/1 drm/i915: Add hot_plug hook for hdmi encoder

This patch adds a separate probe function for HDMI EDID read over DDC channel. This function has been registered as a .hot_plug handler for HDMI encoder.

The current implementation of hdmi_detect()
function re-sets the cached HDMI edid (in connector->detect_edid) in every detect call.This function gets called many times, sometimes directly from userspace probes, forcing drivers to read EDID every detect function call.This causes several problems like:

1. Race conditions in multiple hot_plug / unplug cases, between interrupts bottom halves and userspace detections.
2. Many Un-necessary EDID reads for single hotplug/unplug
3. HDMI complaince failures which expects only one EDID read per hotplug

This function will be serving the purpose of really reading the EDID by really probing the DDC channel, and updating the cached EDID.

The plan is to:
1. i915 IRQ handler bottom half function already calls intel_encoder->hotplug() function. Adding This probe function which will read the EDID only in case of a hotplug / unplug.
2. During init_connector this probe will be called to read the edid
3. Reuse the cached EDID in hdmi_detect() function.

The "< gen7" check is there because this was tested only for >=gen7 platforms. For older platforms the hotplug/reading edid path remains same.

v2: Calling set_edid instead of hdmi_probe during init.
Also, for platforms having DDI, intel_encoder for DP and HDMI is same (taken from intel_dig_port), so for DP also, hot_plug function gets called which is not intended here. So, check for HDMI in intel_hdmi_probe Rely on HPD for updating edid only for platforms gen > 8 and also for VLV.

v3: Dropping the gen < 8 || !VLV check. Now all platforms should rely on hotplug or init for updating the edid.(Daniel)
Also, calling hdmi_probe in init instead of set_edid

v4: Renaming intel_hdmi_probe to intel_hdmi_hot_plug and changing the patch subject. Also calling this hotplug handler from intel_hpd_init to take care of init resume scenarios.

---
 drivers/gpu/drm/i915/intel_hdmi.c    |   54 +++++++++++++++++++++++++++-------
 drivers/gpu/drm/i915/intel_hotplug.c |   11 +++++++
 2 files changed, 54 insertions(+), 11 deletions(-)
drivers/gpu/drm/i915/intel_hdmi.c +43 -11
index bb33c66..9c1a308 100644
@@ -1369,18 +1369,16 @@ intel_hdmi_set_edid(struct drm_connector *connector, bool force)
 	return connected;
 }
 
-static enum drm_connector_status
-intel_hdmi_detect(struct drm_connector *connector, bool force)
+void intel_hdmi_hot_plug(struct intel_encoder *intel_encoder)
 {
-	enum drm_connector_status status;
-	struct intel_hdmi *intel_hdmi = intel_attached_hdmi(connector);
-	struct drm_i915_private *dev_priv = to_i915(connector->dev);
+	struct intel_hdmi *intel_hdmi =
+			enc_to_intel_hdmi(&intel_encoder->base);
+	struct intel_connector *intel_connector =
+				intel_hdmi->attached_connector;
+	struct drm_i915_private *dev_priv = to_i915(intel_encoder->base.dev);
 	bool live_status = false;
 	unsigned int retry = 3;
 
-	DRM_DEBUG_KMS("[CONNECTOR:%d:%s]\n",
-		      connector->base.id, connector->name);
-
 	while (!live_status && --retry) {
 		live_status = intel_digital_port_connected(dev_priv,
 				hdmi_to_dig_port(intel_hdmi));
@@ -1390,15 +1388,48 @@ intel_hdmi_detect(struct drm_connector *connector, bool force)
 	if (!live_status)
 		DRM_DEBUG_KMS("Live status not up!");
 
-	intel_hdmi_unset_edid(connector);
+	/*
+	 * We are here, means there is a hotplug or a force
+	 * detection. Clear the cached EDID and probe the
+	 * DDC bus to check the current status of HDMI.
+	 */
+	intel_hdmi_unset_edid(&intel_connector->base);
+	if (intel_hdmi_set_edid(&intel_connector->base, live_status))
+		DRM_DEBUG_DRIVER("DDC probe: got EDID\n");
+	else
+		DRM_DEBUG_DRIVER("DDC probe: no EDID\n");
+}
 
-	if (intel_hdmi_set_edid(connector, live_status)) {
+static enum drm_connector_status
+intel_hdmi_detect(struct drm_connector *connector, bool force)
+{
+	enum drm_connector_status status;
+	struct intel_connector *intel_connector =
+				to_intel_connector(connector);
+
+	DRM_DEBUG_KMS("[CONNECTOR:%d:%s]\n",
+		      connector->base.id, connector->name);
+
+	/*
+	 * There are many userspace calls which probe EDID from
+	 * detect path. In case of multiple hotplug/unplug, these
+	 * can cause race conditions while probing EDID. Also its
+	 * waste of CPU cycles to read the EDID again and again
+	 * unless there is a real hotplug.
+	 * So, rely on hotplugs and init to read edid.
+	 * Check connector status based on availability of cached EDID.
+	 */
+
+	if (intel_connector->detect_edid) {
 		struct intel_hdmi *intel_hdmi = intel_attached_hdmi(connector);
 
 		hdmi_to_dig_port(intel_hdmi)->base.type = INTEL_OUTPUT_HDMI;
 		status = connector_status_connected;
-	} else
+		DRM_DEBUG_DRIVER("hdmi status = connected\n");
+	} else {
 		status = connector_status_disconnected;
+		DRM_DEBUG_DRIVER("hdmi status = disconnected\n");
+	}
 
 	return status;
 }
@@ -2131,6 +2162,7 @@ void intel_hdmi_init_connector(struct intel_digital_port *intel_dig_port,
 	intel_connector->unregister = intel_connector_unregister;
 
 	intel_hdmi_add_properties(intel_hdmi, connector);
+	intel_encoder->hot_plug = intel_hdmi_hot_plug;
 
 	intel_connector_attach_encoder(intel_connector, intel_encoder);
 	drm_connector_register(connector);
drivers/gpu/drm/i915/intel_hotplug.c +11 -0
index 53c0173..36e16f6 100644
@@ -459,6 +459,7 @@ void intel_hpd_init(struct drm_i915_private *dev_priv)
 	struct drm_device *dev = dev_priv->dev;
 	struct drm_mode_config *mode_config = &dev->mode_config;
 	struct drm_connector *connector;
+	struct intel_encoder *encoder;
 	int i;
 
 	for_each_hpd_pin(i) {
@@ -482,6 +483,16 @@ void intel_hpd_init(struct drm_i915_private *dev_priv)
 	if (dev_priv->display.hpd_irq_setup)
 		dev_priv->display.hpd_irq_setup(dev);
 	spin_unlock_irq(&dev_priv->irq_lock);
+
+	/*
+	 * Connected boot / resume scenarios can't generate new hot plug.
+	 * So, probe it manually.
+	 */
+	list_for_each_entry(encoder, &dev->mode_config.encoder_list,
+			    base.head) {
+		if (encoder->hot_plug)
+			encoder->hot_plug(encoder);
+	}
 }
 
 void intel_hpd_init_work(struct drm_i915_private *dev_priv)
Signature · 6 lines
1.7.10.4

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

Re: [PATCH] drm/i915: Add hot_plug hook for hdmi encoder

Quoted 154 lines · Sonika Jindal

This patch adds a separate probe function for HDMI EDID read over DDC channel. This function has been registered as a .hot_plug handler for HDMI encoder.

The current implementation of hdmi_detect()
function re-sets the cached HDMI edid (in connector->detect_edid) in every detect call.This function gets called many times, sometimes directly from userspace probes, forcing drivers to read EDID every detect function call.This causes several problems like:

1. Race conditions in multiple hot_plug / unplug cases, between interrupts bottom halves and userspace detections.
2. Many Un-necessary EDID reads for single hotplug/unplug
3. HDMI complaince failures which expects only one EDID read per hotplug

This function will be serving the purpose of really reading the EDID by really probing the DDC channel, and updating the cached EDID.

The plan is to:
1. i915 IRQ handler bottom half function already calls intel_encoder->hotplug() function. Adding This probe function which will read the EDID only in case of a hotplug / unplug.
2. During init_connector this probe will be called to read the edid
3. Reuse the cached EDID in hdmi_detect() function.

The "< gen7" check is there because this was tested only for >=gen7 platforms. For older platforms the hotplug/reading edid path remains same.

v2: Calling set_edid instead of hdmi_probe during init.
Also, for platforms having DDI, intel_encoder for DP and HDMI is same (taken from intel_dig_port), so for DP also, hot_plug function gets called which is not intended here. So, check for HDMI in intel_hdmi_probe Rely on HPD for updating edid only for platforms gen > 8 and also for VLV.

v3: Dropping the gen < 8 || !VLV check. Now all platforms should rely on hotplug or init for updating the edid.(Daniel)
Also, calling hdmi_probe in init instead of set_edid

v4: Renaming intel_hdmi_probe to intel_hdmi_hot_plug and changing the patch subject. Also calling this hotplug handler from intel_hpd_init to take care of init resume scenarios.

---
 drivers/gpu/drm/i915/intel_hdmi.c    |   54 +++++++++++++++++++++++++++-------
 drivers/gpu/drm/i915/intel_hotplug.c |   11 +++++++
 2 files changed, 54 insertions(+), 11 deletions(-)
drivers/gpu/drm/i915/intel_hdmi.c +43 -11
index bb33c66..9c1a308 100644
@@ -1369,18 +1369,16 @@ intel_hdmi_set_edid(struct drm_connector *connector, bool force)
 	return connected;
 }
 
-static enum drm_connector_status
-intel_hdmi_detect(struct drm_connector *connector, bool force)
+void intel_hdmi_hot_plug(struct intel_encoder *intel_encoder)
 {
-	enum drm_connector_status status;
-	struct intel_hdmi *intel_hdmi = intel_attached_hdmi(connector);
-	struct drm_i915_private *dev_priv = to_i915(connector->dev);
+	struct intel_hdmi *intel_hdmi =
+			enc_to_intel_hdmi(&intel_encoder->base);
+	struct intel_connector *intel_connector =
+				intel_hdmi->attached_connector;
+	struct drm_i915_private *dev_priv = to_i915(intel_encoder->base.dev);
 	bool live_status = false;
 	unsigned int retry = 3;
 
-	DRM_DEBUG_KMS("[CONNECTOR:%d:%s]\n",
-		      connector->base.id, connector->name);
-
 	while (!live_status && --retry) {
 		live_status = intel_digital_port_connected(dev_priv,
 				hdmi_to_dig_port(intel_hdmi));
@@ -1390,15 +1388,48 @@ intel_hdmi_detect(struct drm_connector *connector, bool force)
 	if (!live_status)
 		DRM_DEBUG_KMS("Live status not up!");
 
-	intel_hdmi_unset_edid(connector);
+	/*
+	 * We are here, means there is a hotplug or a force
+	 * detection. Clear the cached EDID and probe the
+	 * DDC bus to check the current status of HDMI.
+	 */
+	intel_hdmi_unset_edid(&intel_connector->base);
+	if (intel_hdmi_set_edid(&intel_connector->base, live_status))
+		DRM_DEBUG_DRIVER("DDC probe: got EDID\n");
+	else
+		DRM_DEBUG_DRIVER("DDC probe: no EDID\n");
+}
 
-	if (intel_hdmi_set_edid(connector, live_status)) {
+static enum drm_connector_status
+intel_hdmi_detect(struct drm_connector *connector, bool force)
+{
+	enum drm_connector_status status;
+	struct intel_connector *intel_connector =
+				to_intel_connector(connector);
+
+	DRM_DEBUG_KMS("[CONNECTOR:%d:%s]\n",
+		      connector->base.id, connector->name);
+
+	/*
+	 * There are many userspace calls which probe EDID from
+	 * detect path. In case of multiple hotplug/unplug, these
+	 * can cause race conditions while probing EDID. Also its
+	 * waste of CPU cycles to read the EDID again and again
+	 * unless there is a real hotplug.
+	 * So, rely on hotplugs and init to read edid.
+	 * Check connector status based on availability of cached EDID.
+	 */
+
+	if (intel_connector->detect_edid) {
 		struct intel_hdmi *intel_hdmi = intel_attached_hdmi(connector);
 
 		hdmi_to_dig_port(intel_hdmi)->base.type = INTEL_OUTPUT_HDMI;
 		status = connector_status_connected;
-	} else
+		DRM_DEBUG_DRIVER("hdmi status = connected\n");
+	} else {
 		status = connector_status_disconnected;
+		DRM_DEBUG_DRIVER("hdmi status = disconnected\n");
+	}
 
 	return status;
 }
@@ -2131,6 +2162,7 @@ void intel_hdmi_init_connector(struct intel_digital_port *intel_dig_port,
 	intel_connector->unregister = intel_connector_unregister;
 
 	intel_hdmi_add_properties(intel_hdmi, connector);
+	intel_encoder->hot_plug = intel_hdmi_hot_plug;
 
 	intel_connector_attach_encoder(intel_connector, intel_encoder);
 	drm_connector_register(connector);
drivers/gpu/drm/i915/intel_hotplug.c +1 -0
index 53c0173..36e16f6 100644
@@ -459,6 +459,7 @@ void intel_hpd_init(struct drm_i915_private *dev_priv)
 	struct drm_device *dev = dev_priv->dev;
 	struct drm_mode_config *mode_config = &dev->mode_config;
 	struct drm_connector *connector;
+	struct intel_encoder *encoder;
 	int i;
 
 	for_each_hpd_pin(i) {
@@ -482,6 +483,16 @@ void intel_hpd_init(struct drm_i915_private *dev_priv)
 	if (dev_priv->display.hpd_irq_setup)
 		dev_priv->display.hpd_irq_setup(dev);
 	spin_unlock_irq(&dev_priv->irq_lock);

drivers/gpu/drm/i915/intel_hotplug.c void intel_hpd_init(struct drm_i915_private *dev_priv) · line 495

Quoted 10 lines
drivers/gpu/drm/i915/intel_hotplug.c +10 -0
+
+	/*
+	 * Connected boot / resume scenarios can't generate new hot plug.
+	 * So, probe it manually.
+	 */
+	list_for_each_entry(encoder, &dev->mode_config.encoder_list,
+			    base.head) {
+		if (encoder->hot_plug)
+			encoder->hot_plug(encoder);
+	}

Since this doesn't just change hdmi but every encoder I think we should pull it out into a separate prep patch. Especially interactions with dp mst handling look tricky, have you tested that nothing blows up with that? The split-out patch should definitely explain in detail how this interacts with dp hpd handling.
-Daniel

Signature · 7 lines
Daniel Vetter
Software Engineer, Intel Corporation
http://blog.ffwll.ch
_______________________________________________
Intel-gfx mailing list
[email protected]
http://lists.freedesktop.org/mailman/listinfo/intel-gfx
#14 lore raw

Re: [PATCH] drm/i915: Add hot_plug hook for hdmi encoder

Quoted 164 lines · Daniel Vetter
Quoted 163 lines · Sonika Jindal

This patch adds a separate probe function for HDMI EDID read over DDC channel. This function has been registered as a .hot_plug handler for HDMI encoder.

The current implementation of hdmi_detect()
function re-sets the cached HDMI edid (in connector->detect_edid) in every detect call.This function gets called many times, sometimes directly from userspace probes, forcing drivers to read EDID every detect function call.This causes several problems like:

1. Race conditions in multiple hot_plug / unplug cases, between interrupts bottom halves and userspace detections.
2. Many Un-necessary EDID reads for single hotplug/unplug
3. HDMI complaince failures which expects only one EDID read per hotplug

This function will be serving the purpose of really reading the EDID by really probing the DDC channel, and updating the cached EDID.

The plan is to:
1. i915 IRQ handler bottom half function already calls intel_encoder->hotplug() function. Adding This probe function which will read the EDID only in case of a hotplug / unplug.
2. During init_connector this probe will be called to read the edid
3. Reuse the cached EDID in hdmi_detect() function.

The "< gen7" check is there because this was tested only for >=gen7 platforms. For older platforms the hotplug/reading edid path remains same.

v2: Calling set_edid instead of hdmi_probe during init.
Also, for platforms having DDI, intel_encoder for DP and HDMI is same (taken from intel_dig_port), so for DP also, hot_plug function gets called which is not intended here. So, check for HDMI in intel_hdmi_probe Rely on HPD for updating edid only for platforms gen > 8 and also for VLV.

v3: Dropping the gen < 8 || !VLV check. Now all platforms should rely on hotplug or init for updating the edid.(Daniel)
Also, calling hdmi_probe in init instead of set_edid

v4: Renaming intel_hdmi_probe to intel_hdmi_hot_plug and changing the patch subject. Also calling this hotplug handler from intel_hpd_init to take care of init resume scenarios.

---
  drivers/gpu/drm/i915/intel_hdmi.c    |   54 +++++++++++++++++++++++++++-------
  drivers/gpu/drm/i915/intel_hotplug.c |   11 +++++++
  2 files changed, 54 insertions(+), 11 deletions(-)
drivers/gpu/drm/i915/intel_hdmi.c +43 -11
index bb33c66..9c1a308 100644
@@ -1369,18 +1369,16 @@ intel_hdmi_set_edid(struct drm_connector *connector, bool force)
  	return connected;
  }
 
-static enum drm_connector_status
-intel_hdmi_detect(struct drm_connector *connector, bool force)
+void intel_hdmi_hot_plug(struct intel_encoder *intel_encoder)
  {
-	enum drm_connector_status status;
-	struct intel_hdmi *intel_hdmi = intel_attached_hdmi(connector);
-	struct drm_i915_private *dev_priv = to_i915(connector->dev);
+	struct intel_hdmi *intel_hdmi =
+			enc_to_intel_hdmi(&intel_encoder->base);
+	struct intel_connector *intel_connector =
+				intel_hdmi->attached_connector;
+	struct drm_i915_private *dev_priv = to_i915(intel_encoder->base.dev);
  	bool live_status = false;
  	unsigned int retry = 3;
 
-	DRM_DEBUG_KMS("[CONNECTOR:%d:%s]\n",
-		      connector->base.id, connector->name);
-
  	while (!live_status && --retry) {
  		live_status = intel_digital_port_connected(dev_priv,
  				hdmi_to_dig_port(intel_hdmi));
@@ -1390,15 +1388,48 @@ intel_hdmi_detect(struct drm_connector *connector, bool force)
  	if (!live_status)
  		DRM_DEBUG_KMS("Live status not up!");
 
-	intel_hdmi_unset_edid(connector);
+	/*
+	 * We are here, means there is a hotplug or a force
+	 * detection. Clear the cached EDID and probe the
+	 * DDC bus to check the current status of HDMI.
+	 */
+	intel_hdmi_unset_edid(&intel_connector->base);
+	if (intel_hdmi_set_edid(&intel_connector->base, live_status))
+		DRM_DEBUG_DRIVER("DDC probe: got EDID\n");
+	else
+		DRM_DEBUG_DRIVER("DDC probe: no EDID\n");
+}
 
-	if (intel_hdmi_set_edid(connector, live_status)) {
+static enum drm_connector_status
+intel_hdmi_detect(struct drm_connector *connector, bool force)
+{
+	enum drm_connector_status status;
+	struct intel_connector *intel_connector =
+				to_intel_connector(connector);
+
+	DRM_DEBUG_KMS("[CONNECTOR:%d:%s]\n",
+		      connector->base.id, connector->name);
+
+	/*
+	 * There are many userspace calls which probe EDID from
+	 * detect path. In case of multiple hotplug/unplug, these
+	 * can cause race conditions while probing EDID. Also its
+	 * waste of CPU cycles to read the EDID again and again
+	 * unless there is a real hotplug.
+	 * So, rely on hotplugs and init to read edid.
+	 * Check connector status based on availability of cached EDID.
+	 */
+
+	if (intel_connector->detect_edid) {
  		struct intel_hdmi *intel_hdmi = intel_attached_hdmi(connector);
 
  		hdmi_to_dig_port(intel_hdmi)->base.type = INTEL_OUTPUT_HDMI;
  		status = connector_status_connected;
-	} else
+		DRM_DEBUG_DRIVER("hdmi status = connected\n");
+	} else {
  		status = connector_status_disconnected;
+		DRM_DEBUG_DRIVER("hdmi status = disconnected\n");
+	}
 
  	return status;
  }
@@ -2131,6 +2162,7 @@ void intel_hdmi_init_connector(struct intel_digital_port *intel_dig_port,
  	intel_connector->unregister = intel_connector_unregister;
 
  	intel_hdmi_add_properties(intel_hdmi, connector);
+	intel_encoder->hot_plug = intel_hdmi_hot_plug;
 
  	intel_connector_attach_encoder(intel_connector, intel_encoder);
  	drm_connector_register(connector);
drivers/gpu/drm/i915/intel_hotplug.c +10 -0
index 53c0173..36e16f6 100644
@@ -459,6 +459,7 @@ void intel_hpd_init(struct drm_i915_private *dev_priv)
  	struct drm_device *dev = dev_priv->dev;
  	struct drm_mode_config *mode_config = &dev->mode_config;
  	struct drm_connector *connector;
+	struct intel_encoder *encoder;
  	int i;
 
  	for_each_hpd_pin(i) {
@@ -482,6 +483,16 @@ void intel_hpd_init(struct drm_i915_private *dev_priv)
  	if (dev_priv->display.hpd_irq_setup)
  		dev_priv->display.hpd_irq_setup(dev);
  	spin_unlock_irq(&dev_priv->irq_lock);
+
+	/*
+	 * Connected boot / resume scenarios can't generate new hot plug.
+	 * So, probe it manually.
+	 */
+	list_for_each_entry(encoder, &dev->mode_config.encoder_list,
+			    base.head) {
+		if (encoder->hot_plug)
+			encoder->hot_plug(encoder);
Quoted 9 lines

drivers/gpu/drm/i915/intel_hotplug.c void intel_hpd_init(struct drm_i915_private *dev_priv) · line 495

Quoted 1 line
drivers/gpu/drm/i915/intel_hotplug.c +1 -0
+	}

Since this doesn't just change hdmi but every encoder I think we should pull it out into a separate prep patch. Especially interactions with dp mst handling look tricky, have you tested that nothing blows up with that? The split-out patch should definitely explain in detail how this interacts with dp hpd handling.
-Daniel

Right now we don't set hot_plug for DP/edp encoder.
It was used only for sdvo and with this patch we add for hdmi.
Also, in sdvo_init right after assigning the hot_plug function pointer, we call its hot_plug function for initialization. Maybe with this patch, we can remove that part also from sdvo_init.

I can split this patch but without this change, it will result in regression with just one patch. So I thought it would be better to keep them together in one patch.

Regards,
Sonika

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

#15 lore raw

Re: [PATCH] drm/i915: Add hot_plug hook for hdmi encoder

Quoted 176 lines · Sonika
Quoted 164 lines · Daniel Vetter
Quoted 163 lines · Sonika Jindal

This patch adds a separate probe function for HDMI EDID read over DDC channel. This function has been registered as a .hot_plug handler for HDMI encoder.

The current implementation of hdmi_detect()
function re-sets the cached HDMI edid (in connector->detect_edid) in every detect call.This function gets called many times, sometimes directly from userspace probes, forcing drivers to read EDID every detect function call.This causes several problems like:

1. Race conditions in multiple hot_plug / unplug cases, between interrupts bottom halves and userspace detections.
2. Many Un-necessary EDID reads for single hotplug/unplug
3. HDMI complaince failures which expects only one EDID read per hotplug

This function will be serving the purpose of really reading the EDID by really probing the DDC channel, and updating the cached EDID.

The plan is to:
1. i915 IRQ handler bottom half function already calls intel_encoder->hotplug() function. Adding This probe function which will read the EDID only in case of a hotplug / unplug.
2. During init_connector this probe will be called to read the edid
3. Reuse the cached EDID in hdmi_detect() function.

The "< gen7" check is there because this was tested only for >=gen7 platforms. For older platforms the hotplug/reading edid path remains same.

v2: Calling set_edid instead of hdmi_probe during init.
Also, for platforms having DDI, intel_encoder for DP and HDMI is same (taken from intel_dig_port), so for DP also, hot_plug function gets called which is not intended here. So, check for HDMI in intel_hdmi_probe Rely on HPD for updating edid only for platforms gen > 8 and also for VLV.

v3: Dropping the gen < 8 || !VLV check. Now all platforms should rely on hotplug or init for updating the edid.(Daniel)
Also, calling hdmi_probe in init instead of set_edid

v4: Renaming intel_hdmi_probe to intel_hdmi_hot_plug and changing the patch subject. Also calling this hotplug handler from intel_hpd_init to take care of init resume scenarios.

---
 drivers/gpu/drm/i915/intel_hdmi.c    |   54 +++++++++++++++++++++++++++-------
 drivers/gpu/drm/i915/intel_hotplug.c |   11 +++++++
 2 files changed, 54 insertions(+), 11 deletions(-)
drivers/gpu/drm/i915/intel_hdmi.c +43 -11
index bb33c66..9c1a308 100644
@@ -1369,18 +1369,16 @@ intel_hdmi_set_edid(struct drm_connector *connector, bool force)
 	return connected;
 }
 
-static enum drm_connector_status
-intel_hdmi_detect(struct drm_connector *connector, bool force)
+void intel_hdmi_hot_plug(struct intel_encoder *intel_encoder)
 {
-	enum drm_connector_status status;
-	struct intel_hdmi *intel_hdmi = intel_attached_hdmi(connector);
-	struct drm_i915_private *dev_priv = to_i915(connector->dev);
+	struct intel_hdmi *intel_hdmi =
+			enc_to_intel_hdmi(&intel_encoder->base);
+	struct intel_connector *intel_connector =
+				intel_hdmi->attached_connector;
+	struct drm_i915_private *dev_priv = to_i915(intel_encoder->base.dev);
 	bool live_status = false;
 	unsigned int retry = 3;
 
-	DRM_DEBUG_KMS("[CONNECTOR:%d:%s]\n",
-		      connector->base.id, connector->name);
-
 	while (!live_status && --retry) {
 		live_status = intel_digital_port_connected(dev_priv,
 				hdmi_to_dig_port(intel_hdmi));
@@ -1390,15 +1388,48 @@ intel_hdmi_detect(struct drm_connector *connector, bool force)
 	if (!live_status)
 		DRM_DEBUG_KMS("Live status not up!");
 
-	intel_hdmi_unset_edid(connector);
+	/*
+	 * We are here, means there is a hotplug or a force
+	 * detection. Clear the cached EDID and probe the
+	 * DDC bus to check the current status of HDMI.
+	 */
+	intel_hdmi_unset_edid(&intel_connector->base);
+	if (intel_hdmi_set_edid(&intel_connector->base, live_status))
+		DRM_DEBUG_DRIVER("DDC probe: got EDID\n");
+	else
+		DRM_DEBUG_DRIVER("DDC probe: no EDID\n");
+}
 
-	if (intel_hdmi_set_edid(connector, live_status)) {
+static enum drm_connector_status
+intel_hdmi_detect(struct drm_connector *connector, bool force)
+{
+	enum drm_connector_status status;
+	struct intel_connector *intel_connector =
+				to_intel_connector(connector);
+
+	DRM_DEBUG_KMS("[CONNECTOR:%d:%s]\n",
+		      connector->base.id, connector->name);
+
+	/*
+	 * There are many userspace calls which probe EDID from
+	 * detect path. In case of multiple hotplug/unplug, these
+	 * can cause race conditions while probing EDID. Also its
+	 * waste of CPU cycles to read the EDID again and again
+	 * unless there is a real hotplug.
+	 * So, rely on hotplugs and init to read edid.
+	 * Check connector status based on availability of cached EDID.
+	 */
+
+	if (intel_connector->detect_edid) {
 		struct intel_hdmi *intel_hdmi = intel_attached_hdmi(connector);
 
 		hdmi_to_dig_port(intel_hdmi)->base.type = INTEL_OUTPUT_HDMI;
 		status = connector_status_connected;
-	} else
+		DRM_DEBUG_DRIVER("hdmi status = connected\n");
+	} else {
 		status = connector_status_disconnected;
+		DRM_DEBUG_DRIVER("hdmi status = disconnected\n");
+	}
 
 	return status;
 }
@@ -2131,6 +2162,7 @@ void intel_hdmi_init_connector(struct intel_digital_port *intel_dig_port,
 	intel_connector->unregister = intel_connector_unregister;
 
 	intel_hdmi_add_properties(intel_hdmi, connector);
+	intel_encoder->hot_plug = intel_hdmi_hot_plug;
 
 	intel_connector_attach_encoder(intel_connector, intel_encoder);
 	drm_connector_register(connector);
drivers/gpu/drm/i915/intel_hotplug.c +10 -0
index 53c0173..36e16f6 100644
@@ -459,6 +459,7 @@ void intel_hpd_init(struct drm_i915_private *dev_priv)
 	struct drm_device *dev = dev_priv->dev;
 	struct drm_mode_config *mode_config = &dev->mode_config;
 	struct drm_connector *connector;
+	struct intel_encoder *encoder;
 	int i;
 
 	for_each_hpd_pin(i) {
@@ -482,6 +483,16 @@ void intel_hpd_init(struct drm_i915_private *dev_priv)
 	if (dev_priv->display.hpd_irq_setup)
 		dev_priv->display.hpd_irq_setup(dev);
 	spin_unlock_irq(&dev_priv->irq_lock);
+
+	/*
+	 * Connected boot / resume scenarios can't generate new hot plug.
+	 * So, probe it manually.
+	 */
+	list_for_each_entry(encoder, &dev->mode_config.encoder_list,
+			    base.head) {
+		if (encoder->hot_plug)
+			encoder->hot_plug(encoder);
Quoted 9 lines

drivers/gpu/drm/i915/intel_hotplug.c void intel_hpd_init(struct drm_i915_private *dev_priv) · line 495

Quoted 1 line
drivers/gpu/drm/i915/intel_hotplug.c +1 -0
+	}

Since this doesn't just change hdmi but every encoder I think we should pull it out into a separate prep patch. Especially interactions with dp mst handling look tricky, have you tested that nothing blows up with that? The split-out patch should definitely explain in detail how this interacts with dp hpd handling.
-Daniel

Quoted 5 lines

Right now we don't set hot_plug for DP/edp encoder.
It was used only for sdvo and with this patch we add for hdmi. Also, in sdvo_init right after assigning the hot_plug function pointer, we call its hot_plug function for initialization.
Maybe with this patch, we can remove that part also from sdvo_init.

Hm right, I forgot that we have too many hotplug callbacks in our driver ;-)

Quoted 3 lines

I can split this patch but without this change, it will result in regression with just one patch. So I thought it would be better to keep them together in one patch.

If you split it out and make it patch 1 and then the hdmi changes patch 2 nothing breaks in between.
-Daniel

Signature · 7 lines
Daniel Vetter
Software Engineer, Intel Corporation
http://blog.ffwll.ch
_______________________________________________
Intel-gfx mailing list
[email protected]
http://lists.freedesktop.org/mailman/listinfo/intel-gfx
#16 lore raw

PATCH 1/2 drm/i915: Call encoder hotplug for init and resume cases

In reply to Daniel Vetter

For all the encoders, call the hot_plug if it is registered. This is required for connected boot and resume cases to generate fake hpd resulting in reading of edid.
Removing the initial sdvo hot_plug call too so that it will be called just once from this loop.

---
 drivers/gpu/drm/i915/intel_hotplug.c |   11 +++++++++++
 drivers/gpu/drm/i915/intel_sdvo.c    |    1 -
 2 files changed, 11 insertions(+), 1 deletion(-)
drivers/gpu/drm/i915/intel_hotplug.c +11 -0
index 53c0173..eac4757 100644
@@ -458,6 +458,7 @@ void intel_hpd_init(struct drm_i915_private *dev_priv)
 {
 	struct drm_device *dev = dev_priv->dev;
 	struct drm_mode_config *mode_config = &dev->mode_config;
+	struct intel_encoder *encoder;
 	struct drm_connector *connector;
 	int i;
 
@@ -482,6 +483,16 @@ void intel_hpd_init(struct drm_i915_private *dev_priv)
 	if (dev_priv->display.hpd_irq_setup)
 		dev_priv->display.hpd_irq_setup(dev);
 	spin_unlock_irq(&dev_priv->irq_lock);
+
+	/*
+	 * Connected boot / resume scenarios can't generate new hot plug.
+	 * So, probe it manually.
+	 */
+	list_for_each_entry(encoder, &dev->mode_config.encoder_list,
+			    base.head) {
+		if (encoder->hot_plug)
+			encoder->hot_plug(encoder);
+	}
 }
 
 void intel_hpd_init_work(struct drm_i915_private *dev_priv)
drivers/gpu/drm/i915/intel_sdvo.c +0 -1
index 05521b5..55859e9 100644
@@ -2433,7 +2433,6 @@ intel_sdvo_dvi_init(struct intel_sdvo *intel_sdvo, int device)
 		 * Ensure that they get re-enabled when an interrupt happens.
 		 */
 		intel_encoder->hot_plug = intel_sdvo_enable_hotplug;
-		intel_sdvo_enable_hotplug(intel_encoder);
 	} else {
 		intel_connector->polled = DRM_CONNECTOR_POLL_CONNECT | DRM_CONNECTOR_POLL_DISCONNECT;
 	}
Signature · 6 lines
1.7.10.4

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

PATCH 2/2 drm/i915: Add hot_plug hook for hdmi encoder

This patch adds a separate probe function for HDMI EDID read over DDC channel. This function has been registered as a .hot_plug handler for HDMI encoder.

The current implementation of hdmi_detect()
function re-sets the cached HDMI edid (in connector->detect_edid) in every detect call.This function gets called many times, sometimes directly from userspace probes, forcing drivers to read EDID every detect function call.This causes several problems like:

1. Race conditions in multiple hot_plug / unplug cases, between interrupts bottom halves and userspace detections.
2. Many Un-necessary EDID reads for single hotplug/unplug
3. HDMI complaince failures which expects only one EDID read per hotplug

This function will be serving the purpose of really reading the EDID by really probing the DDC channel, and updating the cached EDID.

The plan is to:
1. i915 IRQ handler bottom half function already calls intel_encoder->hotplug() function. Adding This probe function which will read the EDID only in case of a hotplug / unplug.
2. During init_connector this probe will be called to read the edid
3. Reuse the cached EDID in hdmi_detect() function.

The "< gen7" check is there because this was tested only for >=gen7 platforms. For older platforms the hotplug/reading edid path remains same.

v2: Calling set_edid instead of hdmi_probe during init.
Also, for platforms having DDI, intel_encoder for DP and HDMI is same (taken from intel_dig_port), so for DP also, hot_plug function gets called which is not intended here. So, check for HDMI in intel_hdmi_probe Rely on HPD for updating edid only for platforms gen > 8 and also for VLV.

v3: Dropping the gen < 8 || !VLV check. Now all platforms should rely on hotplug or init for updating the edid.(Daniel)
Also, calling hdmi_probe in init instead of set_edid

v4: Renaming intel_hdmi_probe to intel_hdmi_hot_plug.
Also calling this hotplug handler from intel_hpd_init to take care of init resume scenarios.

v5: Moved the call to encoder hotplug during init to separate patch(Daniel)

---
 drivers/gpu/drm/i915/intel_hdmi.c |   54 +++++++++++++++++++++++++++++--------
 1 file changed, 43 insertions(+), 11 deletions(-)
drivers/gpu/drm/i915/intel_hdmi.c +43 -11
index bb33c66..9c1a308 100644
@@ -1369,18 +1369,16 @@ intel_hdmi_set_edid(struct drm_connector *connector, bool force)
 	return connected;
 }
 
-static enum drm_connector_status
-intel_hdmi_detect(struct drm_connector *connector, bool force)
+void intel_hdmi_hot_plug(struct intel_encoder *intel_encoder)
 {
-	enum drm_connector_status status;
-	struct intel_hdmi *intel_hdmi = intel_attached_hdmi(connector);
-	struct drm_i915_private *dev_priv = to_i915(connector->dev);
+	struct intel_hdmi *intel_hdmi =
+			enc_to_intel_hdmi(&intel_encoder->base);
+	struct intel_connector *intel_connector =
+				intel_hdmi->attached_connector;
+	struct drm_i915_private *dev_priv = to_i915(intel_encoder->base.dev);
 	bool live_status = false;
 	unsigned int retry = 3;
 
-	DRM_DEBUG_KMS("[CONNECTOR:%d:%s]\n",
-		      connector->base.id, connector->name);
-
 	while (!live_status && --retry) {
 		live_status = intel_digital_port_connected(dev_priv,
 				hdmi_to_dig_port(intel_hdmi));
@@ -1390,15 +1388,48 @@ intel_hdmi_detect(struct drm_connector *connector, bool force)
 	if (!live_status)
 		DRM_DEBUG_KMS("Live status not up!");
 
-	intel_hdmi_unset_edid(connector);
+	/*
+	 * We are here, means there is a hotplug or a force
+	 * detection. Clear the cached EDID and probe the
+	 * DDC bus to check the current status of HDMI.
+	 */
+	intel_hdmi_unset_edid(&intel_connector->base);
+	if (intel_hdmi_set_edid(&intel_connector->base, live_status))
+		DRM_DEBUG_DRIVER("DDC probe: got EDID\n");
+	else
+		DRM_DEBUG_DRIVER("DDC probe: no EDID\n");
+}
 
-	if (intel_hdmi_set_edid(connector, live_status)) {
+static enum drm_connector_status
+intel_hdmi_detect(struct drm_connector *connector, bool force)
+{
+	enum drm_connector_status status;
+	struct intel_connector *intel_connector =
+				to_intel_connector(connector);
+
+	DRM_DEBUG_KMS("[CONNECTOR:%d:%s]\n",
+		      connector->base.id, connector->name);
+
+	/*
+	 * There are many userspace calls which probe EDID from
+	 * detect path. In case of multiple hotplug/unplug, these
+	 * can cause race conditions while probing EDID. Also its
+	 * waste of CPU cycles to read the EDID again and again
+	 * unless there is a real hotplug.
+	 * So, rely on hotplugs and init to read edid.
+	 * Check connector status based on availability of cached EDID.
+	 */
+
+	if (intel_connector->detect_edid) {
 		struct intel_hdmi *intel_hdmi = intel_attached_hdmi(connector);
 
 		hdmi_to_dig_port(intel_hdmi)->base.type = INTEL_OUTPUT_HDMI;
 		status = connector_status_connected;
-	} else
+		DRM_DEBUG_DRIVER("hdmi status = connected\n");
+	} else {
 		status = connector_status_disconnected;
+		DRM_DEBUG_DRIVER("hdmi status = disconnected\n");
+	}
 
 	return status;
 }
@@ -2131,6 +2162,7 @@ void intel_hdmi_init_connector(struct intel_digital_port *intel_dig_port,
 	intel_connector->unregister = intel_connector_unregister;
 
 	intel_hdmi_add_properties(intel_hdmi, connector);
+	intel_encoder->hot_plug = intel_hdmi_hot_plug;
 
 	intel_connector_attach_encoder(intel_connector, intel_encoder);
 	drm_connector_register(connector);
Signature · 6 lines
1.7.10.4

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

Re: [PATCH 2/2] drm/i915: Add hot_plug hook for hdmi encoder

Quoted 38 lines · Sonika Jindal

This patch adds a separate probe function for HDMI EDID read over DDC channel. This function has been registered as a .hot_plug handler for HDMI encoder.

The current implementation of hdmi_detect()
function re-sets the cached HDMI edid (in connector->detect_edid) in every detect call.This function gets called many times, sometimes directly from userspace probes, forcing drivers to read EDID every detect function call.This causes several problems like:

1. Race conditions in multiple hot_plug / unplug cases, between interrupts bottom halves and userspace detections.
2. Many Un-necessary EDID reads for single hotplug/unplug
3. HDMI complaince failures which expects only one EDID read per hotplug

This function will be serving the purpose of really reading the EDID by really probing the DDC channel, and updating the cached EDID.

The plan is to:
1. i915 IRQ handler bottom half function already calls intel_encoder->hotplug() function. Adding This probe function which will read the EDID only in case of a hotplug / unplug.
2. During init_connector this probe will be called to read the edid
3. Reuse the cached EDID in hdmi_detect() function.

The "< gen7" check is there because this was tested only for >=gen7 platforms. For older platforms the hotplug/reading edid path remains same.

v2: Calling set_edid instead of hdmi_probe during init.
Also, for platforms having DDI, intel_encoder for DP and HDMI is same (taken from intel_dig_port), so for DP also, hot_plug function gets called which is not intended here. So, check for HDMI in intel_hdmi_probe Rely on HPD for updating edid only for platforms gen > 8 and also for VLV.

v3: Dropping the gen < 8 || !VLV check. Now all platforms should rely on hotplug or init for updating the edid.(Daniel)
Also, calling hdmi_probe in init instead of set_edid

Quoted 8 lines

v4: Renaming intel_hdmi_probe to intel_hdmi_hot_plug.
Also calling this hotplug handler from intel_hpd_init to take care of init resume scenarios.

v5: Moved the call to encoder hotplug during init to separate patch(Daniel)

I merged the first patch in this series, but I have a quick question on this one here. Can you pls ping me on irc?

Thanks, Daniel

Quoted 100 lines
---
 drivers/gpu/drm/i915/intel_hdmi.c |   54 +++++++++++++++++++++++++++++--------
 1 file changed, 43 insertions(+), 11 deletions(-)
drivers/gpu/drm/i915/intel_hdmi.c +43 -11
index bb33c66..9c1a308 100644
@@ -1369,18 +1369,16 @@ intel_hdmi_set_edid(struct drm_connector *connector, bool force)
 	return connected;
 }
 
-static enum drm_connector_status
-intel_hdmi_detect(struct drm_connector *connector, bool force)
+void intel_hdmi_hot_plug(struct intel_encoder *intel_encoder)
 {
-	enum drm_connector_status status;
-	struct intel_hdmi *intel_hdmi = intel_attached_hdmi(connector);
-	struct drm_i915_private *dev_priv = to_i915(connector->dev);
+	struct intel_hdmi *intel_hdmi =
+			enc_to_intel_hdmi(&intel_encoder->base);
+	struct intel_connector *intel_connector =
+				intel_hdmi->attached_connector;
+	struct drm_i915_private *dev_priv = to_i915(intel_encoder->base.dev);
 	bool live_status = false;
 	unsigned int retry = 3;
 
-	DRM_DEBUG_KMS("[CONNECTOR:%d:%s]\n",
-		      connector->base.id, connector->name);
-
 	while (!live_status && --retry) {
 		live_status = intel_digital_port_connected(dev_priv,
 				hdmi_to_dig_port(intel_hdmi));
@@ -1390,15 +1388,48 @@ intel_hdmi_detect(struct drm_connector *connector, bool force)
 	if (!live_status)
 		DRM_DEBUG_KMS("Live status not up!");
 
-	intel_hdmi_unset_edid(connector);
+	/*
+	 * We are here, means there is a hotplug or a force
+	 * detection. Clear the cached EDID and probe the
+	 * DDC bus to check the current status of HDMI.
+	 */
+	intel_hdmi_unset_edid(&intel_connector->base);
+	if (intel_hdmi_set_edid(&intel_connector->base, live_status))
+		DRM_DEBUG_DRIVER("DDC probe: got EDID\n");
+	else
+		DRM_DEBUG_DRIVER("DDC probe: no EDID\n");
+}
 
-	if (intel_hdmi_set_edid(connector, live_status)) {
+static enum drm_connector_status
+intel_hdmi_detect(struct drm_connector *connector, bool force)
+{
+	enum drm_connector_status status;
+	struct intel_connector *intel_connector =
+				to_intel_connector(connector);
+
+	DRM_DEBUG_KMS("[CONNECTOR:%d:%s]\n",
+		      connector->base.id, connector->name);
+
+	/*
+	 * There are many userspace calls which probe EDID from
+	 * detect path. In case of multiple hotplug/unplug, these
+	 * can cause race conditions while probing EDID. Also its
+	 * waste of CPU cycles to read the EDID again and again
+	 * unless there is a real hotplug.
+	 * So, rely on hotplugs and init to read edid.
+	 * Check connector status based on availability of cached EDID.
+	 */
+
+	if (intel_connector->detect_edid) {
 		struct intel_hdmi *intel_hdmi = intel_attached_hdmi(connector);
 
 		hdmi_to_dig_port(intel_hdmi)->base.type = INTEL_OUTPUT_HDMI;
 		status = connector_status_connected;
-	} else
+		DRM_DEBUG_DRIVER("hdmi status = connected\n");
+	} else {
 		status = connector_status_disconnected;
+		DRM_DEBUG_DRIVER("hdmi status = disconnected\n");
+	}
 
 	return status;
 }
@@ -2131,6 +2162,7 @@ void intel_hdmi_init_connector(struct intel_digital_port *intel_dig_port,
 	intel_connector->unregister = intel_connector_unregister;
 
 	intel_hdmi_add_properties(intel_hdmi, connector);
+	intel_encoder->hot_plug = intel_hdmi_hot_plug;
 
 	intel_connector_attach_encoder(intel_connector, intel_encoder);
 	drm_connector_register(connector);
Signature · 6 lines
1.7.10.4

_______________________________________________
Intel-gfx mailing list
[email protected]
http://lists.freedesktop.org/mailman/listinfo/intel-gfx
Signature · 7 lines
Daniel Vetter
Software Engineer, Intel Corporation
http://blog.ffwll.ch
_______________________________________________
Intel-gfx mailing list
[email protected]
http://lists.freedesktop.org/mailman/listinfo/intel-gfx
#19 lore raw

PATCH 2/2 drm/i915: Add hot_plug hook for hdmi encoder

This patch adds a separate probe function for HDMI EDID read over DDC channel. This function has been registered as a .hot_plug handler for HDMI encoder.

The current implementation of hdmi_detect()
function re-sets the cached HDMI edid (in connector->detect_edid) in every detect call.This function gets called many times, sometimes directly from userspace probes, forcing drivers to read EDID every detect function call.This causes several problems like:

1. Race conditions in multiple hot_plug / unplug cases, between interrupts bottom halves and userspace detections.
2. Many Un-necessary EDID reads for single hotplug/unplug
3. HDMI complaince failures which expects only one EDID read per hotplug

This function will be serving the purpose of really reading the EDID by really probing the DDC channel, and updating the cached EDID.

The plan is to:
1. i915 IRQ handler bottom half function already calls intel_encoder->hotplug() function. Adding This probe function which will read the EDID only in case of a hotplug / unplug.
2. During init_connector this probe will be called to read the edid
3. Reuse the cached EDID in hdmi_detect() function.

The "< gen7" check is there because this was tested only for >=gen7 platforms. For older platforms the hotplug/reading edid path remains same.

v2: Calling set_edid instead of hdmi_probe during init.
Also, for platforms having DDI, intel_encoder for DP and HDMI is same (taken from intel_dig_port), so for DP also, hot_plug function gets called which is not intended here. So, check for HDMI in intel_hdmi_probe Rely on HPD for updating edid only for platforms gen > 8 and also for VLV.

v3: Dropping the gen < 8 || !VLV check. Now all platforms should rely on hotplug or init for updating the edid.(Daniel)
Also, calling hdmi_probe in init instead of set_edid

v4: Renaming intel_hdmi_probe to intel_hdmi_hot_plug.
Also calling this hotplug handler from intel_hpd_init to take care of init resume scenarios.

v5: Moved the call to encoder hotplug during init to separate patch(Daniel)
v6: Rebased.

---
 drivers/gpu/drm/i915/intel_hdmi.c | 58 ++++++++++++++++++++++++++++++---------
 1 file changed, 45 insertions(+), 13 deletions(-)
drivers/gpu/drm/i915/intel_hdmi.c +45 -13
index bdd462e..16dd2a7 100644
@@ -1381,18 +1381,16 @@ intel_hdmi_set_edid(struct drm_connector *connector, bool force)
 	return connected;
 }
 
-static enum drm_connector_status
-intel_hdmi_detect(struct drm_connector *connector, bool force)
+void intel_hdmi_hot_plug(struct intel_encoder *intel_encoder)
 {
-	enum drm_connector_status status;
-	struct intel_hdmi *intel_hdmi = intel_attached_hdmi(connector);
-	struct drm_i915_private *dev_priv = to_i915(connector->dev);
+	struct intel_hdmi *intel_hdmi =
+			enc_to_intel_hdmi(&intel_encoder->base);
+	struct intel_connector *intel_connector =
+				intel_hdmi->attached_connector;
+	struct drm_i915_private *dev_priv = to_i915(intel_encoder->base.dev);
 	bool live_status = false;
 	unsigned int retry = 3;
 
-	DRM_DEBUG_KMS("[CONNECTOR:%d:%s]\n",
-		      connector->base.id, connector->name);
-
 	intel_display_power_get(dev_priv, POWER_DOMAIN_GMBUS);
 
 	while (!live_status && --retry) {
@@ -1400,21 +1398,53 @@ intel_hdmi_detect(struct drm_connector *connector, bool force)
 				hdmi_to_dig_port(intel_hdmi));
 		mdelay(10);
 	}
+	intel_display_power_put(dev_priv, POWER_DOMAIN_GMBUS);
 
 	if (!live_status)
 		DRM_DEBUG_KMS("Live status not up!");
 
-	intel_hdmi_unset_edid(connector);
+	/*
+	 * We are here, means there is a hotplug or a force
+	 * detection. Clear the cached EDID and probe the
+	 * DDC bus to check the current status of HDMI.
+	 */
+	intel_hdmi_unset_edid(&intel_connector->base);
+	if (intel_hdmi_set_edid(&intel_connector->base, live_status))
+		DRM_DEBUG_DRIVER("DDC probe: got EDID\n");
+	else
+		DRM_DEBUG_DRIVER("DDC probe: no EDID\n");
+}
+
+static enum drm_connector_status
+intel_hdmi_detect(struct drm_connector *connector, bool force)
+{
+	enum drm_connector_status status;
+	struct intel_connector *intel_connector =
+				to_intel_connector(connector);
 
-	if (intel_hdmi_set_edid(connector, live_status)) {
+	DRM_DEBUG_KMS("[CONNECTOR:%d:%s]\n",
+		      connector->base.id, connector->name);
+
+	/*
+	 * There are many userspace calls which probe EDID from
+	 * detect path. In case of multiple hotplug/unplug, these
+	 * can cause race conditions while probing EDID. Also its
+	 * waste of CPU cycles to read the EDID again and again
+	 * unless there is a real hotplug.
+	 * So, rely on hotplugs and init to read edid.
+	 * Check connector status based on availability of cached EDID.
+	 */
+
+	if (intel_connector->detect_edid) {
 		struct intel_hdmi *intel_hdmi = intel_attached_hdmi(connector);
 
 		hdmi_to_dig_port(intel_hdmi)->base.type = INTEL_OUTPUT_HDMI;
 		status = connector_status_connected;
-	} else
+		DRM_DEBUG_DRIVER("hdmi status = connected\n");
+	} else {
 		status = connector_status_disconnected;
-
-	intel_display_power_put(dev_priv, POWER_DOMAIN_GMBUS);
+		DRM_DEBUG_DRIVER("hdmi status = disconnected\n");
+	}
 
 	return status;
 }
@@ -2131,6 +2161,8 @@ void intel_hdmi_init_connector(struct intel_digital_port *intel_dig_port,
 
 	intel_hdmi_add_properties(intel_hdmi, connector);
 
+	intel_encoder->hot_plug = intel_hdmi_hot_plug;
+
 	intel_connector_attach_encoder(intel_connector, intel_encoder);
 	drm_connector_register(connector);
 	intel_hdmi->attached_connector = intel_connector;
Signature · 6 lines
1.9.1

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

PATCH 1/1 drm/i915: Add hot_plug hook for hdmi encoder

From: Shashank Sharma <[email protected]>

This patch adds a separate probe function for HDMI EDID read over DDC channel. This function has been registered as a .hot_plug handler for HDMI encoder.

The current implementation of hdmi_detect()
function re-sets the cached HDMI edid (in connector->detect_edid) in every detect call.This function gets called many times, sometimes directly from userspace probes, forcing drivers to read EDID every detect function call.This causes several problems like:

1. Race conditions in multiple hot_plug / unplug cases, between interrupts bottom halves and userspace detections.
2. Many Un-necessary EDID reads for single hotplug/unplug
3. HDMI complaince failures which expects only one EDID read per hotplug

This function will be serving the purpose of really reading the EDID by really probing the DDC channel, and updating the cached EDID.

The plan is to:
1. i915 IRQ handler bottom half function already calls intel_encoder->hotplug() function. Adding This probe function which will read the EDID only in case of a hotplug / unplug.
2. During init_connector this probe will be called to read the edid
3. Reuse the cached EDID in hdmi_detect() function.

The "< gen7" check is there because this was tested only for >=gen7 platforms. For older platforms the hotplug/reading edid path remains same.

v2: Calling set_edid instead of hdmi_probe during init.
Also, for platforms having DDI, intel_encoder for DP and HDMI is same (taken from intel_dig_port), so for DP also, hot_plug function gets called which is not intended here. So, check for HDMI in intel_hdmi_probe Rely on HPD for updating edid only for platforms gen > 8 and also for VLV.

v3: Dropping the gen < 8 || !VLV check. Now all platforms should rely on hotplug or init for updating the edid.(Daniel)
Also, calling hdmi_probe in init instead of set_edid

v4: Renaming intel_hdmi_probe to intel_hdmi_hot_plug.
Also calling this hotplug handler from intel_hpd_init to take care of init resume scenarios.

v5: Moved the call to encoder hotplug during init to separate patch(Daniel)
v6: Rebased and maintaining authorship.

---
 drivers/gpu/drm/i915/intel_hdmi.c | 58 ++++++++++++++++++++++++++++++---------
 1 file changed, 45 insertions(+), 13 deletions(-)
drivers/gpu/drm/i915/intel_hdmi.c +45 -13
index bdd462e..16dd2a7 100644
@@ -1381,18 +1381,16 @@ intel_hdmi_set_edid(struct drm_connector *connector, bool force)
 	return connected;
 }
 
-static enum drm_connector_status
-intel_hdmi_detect(struct drm_connector *connector, bool force)
+void intel_hdmi_hot_plug(struct intel_encoder *intel_encoder)
 {
-	enum drm_connector_status status;
-	struct intel_hdmi *intel_hdmi = intel_attached_hdmi(connector);
-	struct drm_i915_private *dev_priv = to_i915(connector->dev);
+	struct intel_hdmi *intel_hdmi =
+			enc_to_intel_hdmi(&intel_encoder->base);
+	struct intel_connector *intel_connector =
+				intel_hdmi->attached_connector;
+	struct drm_i915_private *dev_priv = to_i915(intel_encoder->base.dev);
 	bool live_status = false;
 	unsigned int retry = 3;
 
-	DRM_DEBUG_KMS("[CONNECTOR:%d:%s]\n",
-		      connector->base.id, connector->name);
-
 	intel_display_power_get(dev_priv, POWER_DOMAIN_GMBUS);
 
 	while (!live_status && --retry) {
@@ -1400,21 +1398,53 @@ intel_hdmi_detect(struct drm_connector *connector, bool force)
 				hdmi_to_dig_port(intel_hdmi));
 		mdelay(10);
 	}
+	intel_display_power_put(dev_priv, POWER_DOMAIN_GMBUS);
 
 	if (!live_status)
 		DRM_DEBUG_KMS("Live status not up!");
 
-	intel_hdmi_unset_edid(connector);
+	/*
+	 * We are here, means there is a hotplug or a force
+	 * detection. Clear the cached EDID and probe the
+	 * DDC bus to check the current status of HDMI.
+	 */
+	intel_hdmi_unset_edid(&intel_connector->base);
+	if (intel_hdmi_set_edid(&intel_connector->base, live_status))
+		DRM_DEBUG_DRIVER("DDC probe: got EDID\n");
+	else
+		DRM_DEBUG_DRIVER("DDC probe: no EDID\n");
+}
+
+static enum drm_connector_status
+intel_hdmi_detect(struct drm_connector *connector, bool force)
+{
+	enum drm_connector_status status;
+	struct intel_connector *intel_connector =
+				to_intel_connector(connector);
 
-	if (intel_hdmi_set_edid(connector, live_status)) {
+	DRM_DEBUG_KMS("[CONNECTOR:%d:%s]\n",
+		      connector->base.id, connector->name);
+
+	/*
+	 * There are many userspace calls which probe EDID from
+	 * detect path. In case of multiple hotplug/unplug, these
+	 * can cause race conditions while probing EDID. Also its
+	 * waste of CPU cycles to read the EDID again and again
+	 * unless there is a real hotplug.
+	 * So, rely on hotplugs and init to read edid.
+	 * Check connector status based on availability of cached EDID.
+	 */
+
+	if (intel_connector->detect_edid) {
 		struct intel_hdmi *intel_hdmi = intel_attached_hdmi(connector);
 
 		hdmi_to_dig_port(intel_hdmi)->base.type = INTEL_OUTPUT_HDMI;
 		status = connector_status_connected;
-	} else
+		DRM_DEBUG_DRIVER("hdmi status = connected\n");
+	} else {
 		status = connector_status_disconnected;
-
-	intel_display_power_put(dev_priv, POWER_DOMAIN_GMBUS);
+		DRM_DEBUG_DRIVER("hdmi status = disconnected\n");
+	}
 
 	return status;
 }
@@ -2131,6 +2161,8 @@ void intel_hdmi_init_connector(struct intel_digital_port *intel_dig_port,
 
 	intel_hdmi_add_properties(intel_hdmi, connector);
 
+	intel_encoder->hot_plug = intel_hdmi_hot_plug;
+
 	intel_connector_attach_encoder(intel_connector, intel_encoder);
 	drm_connector_register(connector);
 	intel_hdmi->attached_connector = intel_connector;
Signature · 6 lines
1.9.1

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

Re: [PATCH] drm/i915: Add hot_plug hook for hdmi encoder

Quoted 44 lines · Sonika Jindal

From: Shashank Sharma <[email protected]>

This patch adds a separate probe function for HDMI EDID read over DDC channel. This function has been registered as a .hot_plug handler for HDMI encoder.

The current implementation of hdmi_detect()
function re-sets the cached HDMI edid (in connector->detect_edid) in every detect call.This function gets called many times, sometimes directly from userspace probes, forcing drivers to read EDID every detect function call.This causes several problems like:

1. Race conditions in multiple hot_plug / unplug cases, between interrupts bottom halves and userspace detections.
2. Many Un-necessary EDID reads for single hotplug/unplug
3. HDMI complaince failures which expects only one EDID read per hotplug

This function will be serving the purpose of really reading the EDID by really probing the DDC channel, and updating the cached EDID.

The plan is to:
1. i915 IRQ handler bottom half function already calls intel_encoder->hotplug() function. Adding This probe function which will read the EDID only in case of a hotplug / unplug.
2. During init_connector this probe will be called to read the edid
3. Reuse the cached EDID in hdmi_detect() function.

The "< gen7" check is there because this was tested only for >=gen7 platforms. For older platforms the hotplug/reading edid path remains same.

v2: Calling set_edid instead of hdmi_probe during init.
Also, for platforms having DDI, intel_encoder for DP and HDMI is same (taken from intel_dig_port), so for DP also, hot_plug function gets called which is not intended here. So, check for HDMI in intel_hdmi_probe Rely on HPD for updating edid only for platforms gen > 8 and also for VLV.

v3: Dropping the gen < 8 || !VLV check. Now all platforms should rely on hotplug or init for updating the edid.(Daniel)
Also, calling hdmi_probe in init instead of set_edid

v4: Renaming intel_hdmi_probe to intel_hdmi_hot_plug.
Also calling this hotplug handler from intel_hpd_init to take care of init resume scenarios.

Quoted 5 lines

v5: Moved the call to encoder hotplug during init to separate patch(Daniel)
v6: Rebased and maintaining authorship.

Iirc we had to revert an earlier version of this. Commit message should cite that revert and explain why this version is better.

Also after a few days please start a new thread when resubmitting a patch series. The idea behind in-reply-to is to keep the discussion together while it's still ongoing. That's not the case when the discussion has died down since weeks.

Also, your patch to take hpd sense into account for hdmi detection has causes a regression. Your on the cc. Handling that regression needs to be absolute top priority (otherwise I need to revert that patch too).

Thanks, Daniel

Quoted 108 lines
---
 drivers/gpu/drm/i915/intel_hdmi.c | 58 ++++++++++++++++++++++++++++++---------
 1 file changed, 45 insertions(+), 13 deletions(-)
drivers/gpu/drm/i915/intel_hdmi.c +45 -13
index bdd462e..16dd2a7 100644
@@ -1381,18 +1381,16 @@ intel_hdmi_set_edid(struct drm_connector *connector, bool force)
 	return connected;
 }
 
-static enum drm_connector_status
-intel_hdmi_detect(struct drm_connector *connector, bool force)
+void intel_hdmi_hot_plug(struct intel_encoder *intel_encoder)
 {
-	enum drm_connector_status status;
-	struct intel_hdmi *intel_hdmi = intel_attached_hdmi(connector);
-	struct drm_i915_private *dev_priv = to_i915(connector->dev);
+	struct intel_hdmi *intel_hdmi =
+			enc_to_intel_hdmi(&intel_encoder->base);
+	struct intel_connector *intel_connector =
+				intel_hdmi->attached_connector;
+	struct drm_i915_private *dev_priv = to_i915(intel_encoder->base.dev);
 	bool live_status = false;
 	unsigned int retry = 3;
 
-	DRM_DEBUG_KMS("[CONNECTOR:%d:%s]\n",
-		      connector->base.id, connector->name);
-
 	intel_display_power_get(dev_priv, POWER_DOMAIN_GMBUS);
 
 	while (!live_status && --retry) {
@@ -1400,21 +1398,53 @@ intel_hdmi_detect(struct drm_connector *connector, bool force)
 				hdmi_to_dig_port(intel_hdmi));
 		mdelay(10);
 	}
+	intel_display_power_put(dev_priv, POWER_DOMAIN_GMBUS);
 
 	if (!live_status)
 		DRM_DEBUG_KMS("Live status not up!");
 
-	intel_hdmi_unset_edid(connector);
+	/*
+	 * We are here, means there is a hotplug or a force
+	 * detection. Clear the cached EDID and probe the
+	 * DDC bus to check the current status of HDMI.
+	 */
+	intel_hdmi_unset_edid(&intel_connector->base);
+	if (intel_hdmi_set_edid(&intel_connector->base, live_status))
+		DRM_DEBUG_DRIVER("DDC probe: got EDID\n");
+	else
+		DRM_DEBUG_DRIVER("DDC probe: no EDID\n");
+}
+
+static enum drm_connector_status
+intel_hdmi_detect(struct drm_connector *connector, bool force)
+{
+	enum drm_connector_status status;
+	struct intel_connector *intel_connector =
+				to_intel_connector(connector);
 
-	if (intel_hdmi_set_edid(connector, live_status)) {
+	DRM_DEBUG_KMS("[CONNECTOR:%d:%s]\n",
+		      connector->base.id, connector->name);
+
+	/*
+	 * There are many userspace calls which probe EDID from
+	 * detect path. In case of multiple hotplug/unplug, these
+	 * can cause race conditions while probing EDID. Also its
+	 * waste of CPU cycles to read the EDID again and again
+	 * unless there is a real hotplug.
+	 * So, rely on hotplugs and init to read edid.
+	 * Check connector status based on availability of cached EDID.
+	 */
+
+	if (intel_connector->detect_edid) {
 		struct intel_hdmi *intel_hdmi = intel_attached_hdmi(connector);
 
 		hdmi_to_dig_port(intel_hdmi)->base.type = INTEL_OUTPUT_HDMI;
 		status = connector_status_connected;
-	} else
+		DRM_DEBUG_DRIVER("hdmi status = connected\n");
+	} else {
 		status = connector_status_disconnected;
-
-	intel_display_power_put(dev_priv, POWER_DOMAIN_GMBUS);
+		DRM_DEBUG_DRIVER("hdmi status = disconnected\n");
+	}
 
 	return status;
 }
@@ -2131,6 +2161,8 @@ void intel_hdmi_init_connector(struct intel_digital_port *intel_dig_port,
 
 	intel_hdmi_add_properties(intel_hdmi, connector);
 
+	intel_encoder->hot_plug = intel_hdmi_hot_plug;
+
 	intel_connector_attach_encoder(intel_connector, intel_encoder);
 	drm_connector_register(connector);
 	intel_hdmi->attached_connector = intel_connector;
Signature · 6 lines
1.9.1

_______________________________________________
Intel-gfx mailing list
[email protected]
http://lists.freedesktop.org/mailman/listinfo/intel-gfx
Signature · 7 lines
Daniel Vetter
Software Engineer, Intel Corporation
http://blog.ffwll.ch
_______________________________________________
Intel-gfx mailing list
[email protected]
http://lists.freedesktop.org/mailman/listinfo/intel-gfx
#22 lore raw

Re: [PATCH] drm/i915: Add hot_plug hook for hdmi encoder

Sure, I will start a new thread for this.

Regarding the regression, since it is a sandybridge system, I am not sure if I can debug it anyhow because of the unavailability of the system. For the live status check, we earlier suggested to use platforms gen8 and above.
But for wider testing, we removed the check.
If required we can add a check there.

Regards,
Sonika

-----Original Message-----
From: Daniel Vetter [mailto:[email protected]] On Behalf Of Daniel Vetter
Sent: Thursday, December 10, 2015 2:00 PM
To: Jindal, Sonika <[email protected]>
Cc: [email protected]
Subject: Re: [Intel-gfx] [PATCH] drm/i915: Add hot_plug hook for hdmi encoder

Quoted 44 lines · Sonika Jindal

From: Shashank Sharma <[email protected]>

This patch adds a separate probe function for HDMI EDID read over DDC channel. This function has been registered as a .hot_plug handler for HDMI encoder.

The current implementation of hdmi_detect() function re-sets the cached HDMI edid (in connector->detect_edid) in every detect call.This function gets called many times, sometimes directly from userspace probes, forcing drivers to read EDID every detect function call.This causes several problems like:

1. Race conditions in multiple hot_plug / unplug cases, between interrupts bottom halves and userspace detections.
2. Many Un-necessary EDID reads for single hotplug/unplug 3. HDMI complaince failures which expects only one EDID read per hotplug

This function will be serving the purpose of really reading the EDID by really probing the DDC channel, and updating the cached EDID.

The plan is to:
1. i915 IRQ handler bottom half function already calls intel_encoder->hotplug() function. Adding This probe function which will read the EDID only in case of a hotplug / unplug.
2. During init_connector this probe will be called to read the edid 3. Reuse the cached EDID in hdmi_detect() function.

The "< gen7" check is there because this was tested only for >=gen7 platforms. For older platforms the hotplug/reading edid path remains same.

v2: Calling set_edid instead of hdmi_probe during init.
Also, for platforms having DDI, intel_encoder for DP and HDMI is same
(taken from intel_dig_port), so for DP also, hot_plug function gets
called which is not intended here. So, check for HDMI in
intel_hdmi_probe Rely on HPD for updating edid only for platforms gen > 8 and also for VLV.

v3: Dropping the gen < 8 || !VLV check. Now all platforms should rely on hotplug or init for updating the edid.(Daniel) Also, calling hdmi_probe in init instead of set_edid

v4: Renaming intel_hdmi_probe to intel_hdmi_hot_plug.
Also calling this hotplug handler from intel_hpd_init to take care of init resume scenarios.

Quoted 6 lines

v5: Moved the call to encoder hotplug during init to separate patch(Daniel)
v6: Rebased and maintaining authorship.

Iirc we had to revert an earlier version of this. Commit message should cite that revert and explain why this version is better.

Also after a few days please start a new thread when resubmitting a patch series. The idea behind in-reply-to is to keep the discussion together while it's still ongoing. That's not the case when the discussion has died down since weeks.

Also, your patch to take hpd sense into account for hdmi detection has causes a regression. Your on the cc. Handling that regression needs to be absolute top priority (otherwise I need to revert that patch too).

Thanks, Daniel

Quoted 106 lines
---
 drivers/gpu/drm/i915/intel_hdmi.c | 58 
++++++++++++++++++++++++++++++---------
 1 file changed, 45 insertions(+), 13 deletions(-)
drivers/gpu/drm/i915/intel_hdmi.c +44 -13
index bdd462e..16dd2a7 100644
@@ -1381,18 +1381,16 @@ intel_hdmi_set_edid(struct drm_connector *connector, bool force)
 	return connected;
 }
 
-static enum drm_connector_status
-intel_hdmi_detect(struct drm_connector *connector, bool force)
+void intel_hdmi_hot_plug(struct intel_encoder *intel_encoder)
 {
-	enum drm_connector_status status;
-	struct intel_hdmi *intel_hdmi = intel_attached_hdmi(connector);
-	struct drm_i915_private *dev_priv = to_i915(connector->dev);
+	struct intel_hdmi *intel_hdmi =
+			enc_to_intel_hdmi(&intel_encoder->base);
+	struct intel_connector *intel_connector =
+				intel_hdmi->attached_connector;
+	struct drm_i915_private *dev_priv = 
+to_i915(intel_encoder->base.dev);
 	bool live_status = false;
 	unsigned int retry = 3;
 
-	DRM_DEBUG_KMS("[CONNECTOR:%d:%s]\n",
-		      connector->base.id, connector->name);
-
 	intel_display_power_get(dev_priv, POWER_DOMAIN_GMBUS);
 
 	while (!live_status && --retry) {
@@ -1400,21 +1398,53 @@ intel_hdmi_detect(struct drm_connector *connector, bool force)
 				hdmi_to_dig_port(intel_hdmi));
 		mdelay(10);
 	}
+	intel_display_power_put(dev_priv, POWER_DOMAIN_GMBUS);
 
 	if (!live_status)
 		DRM_DEBUG_KMS("Live status not up!");
 
-	intel_hdmi_unset_edid(connector);
+	/*
+	 * We are here, means there is a hotplug or a force
+	 * detection. Clear the cached EDID and probe the
+	 * DDC bus to check the current status of HDMI.
+	 */
+	intel_hdmi_unset_edid(&intel_connector->base);
+	if (intel_hdmi_set_edid(&intel_connector->base, live_status))
+		DRM_DEBUG_DRIVER("DDC probe: got EDID\n");
+	else
+		DRM_DEBUG_DRIVER("DDC probe: no EDID\n"); }
+
+static enum drm_connector_status
+intel_hdmi_detect(struct drm_connector *connector, bool force) {
+	enum drm_connector_status status;
+	struct intel_connector *intel_connector =
+				to_intel_connector(connector);
 
-	if (intel_hdmi_set_edid(connector, live_status)) {
+	DRM_DEBUG_KMS("[CONNECTOR:%d:%s]\n",
+		      connector->base.id, connector->name);
+
+	/*
+	 * There are many userspace calls which probe EDID from
+	 * detect path. In case of multiple hotplug/unplug, these
+	 * can cause race conditions while probing EDID. Also its
+	 * waste of CPU cycles to read the EDID again and again
+	 * unless there is a real hotplug.
+	 * So, rely on hotplugs and init to read edid.
+	 * Check connector status based on availability of cached EDID.
+	 */
+
+	if (intel_connector->detect_edid) {
 		struct intel_hdmi *intel_hdmi = intel_attached_hdmi(connector);
 
 		hdmi_to_dig_port(intel_hdmi)->base.type = INTEL_OUTPUT_HDMI;
 		status = connector_status_connected;
-	} else
+		DRM_DEBUG_DRIVER("hdmi status = connected\n");
+	} else {
 		status = connector_status_disconnected;
-
-	intel_display_power_put(dev_priv, POWER_DOMAIN_GMBUS);
+		DRM_DEBUG_DRIVER("hdmi status = disconnected\n");
+	}
 
 	return status;
 }
@@ -2131,6 +2161,8 @@ void intel_hdmi_init_connector(struct 
 intel_digital_port *intel_dig_port,
 
 	intel_hdmi_add_properties(intel_hdmi, connector);
 
+	intel_encoder->hot_plug = intel_hdmi_hot_plug;
+
 	intel_connector_attach_encoder(intel_connector, intel_encoder);
 	drm_connector_register(connector);

intel_hdmi->attached_connector = intel_connector;
--
1.9.1

Quoted 4 lines

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

--
Daniel Vetter
Software Engineer, Intel Corporation
http://blog.ffwll.ch
_______________________________________________
Intel-gfx mailing list
[email protected]
http://lists.freedesktop.org/mailman/listinfo/intel-gfx

#23 lore raw

Re: [PATCH] drm/i915: Add hot_plug hook for hdmi encoder

Quoted 8 lines · Sonika

Sure, I will start a new thread for this.

Regarding the regression, since it is a sandybridge system, I am not sure if I can debug it anyhow because of the unavailability of the system. For the live status check, we earlier suggested to use platforms gen8 and above.
But for wider testing, we removed the check.
If required we can add a check there.

It's your patch, you need to own this. And until we've confirmed that it's a source issue I won't add a genAnything check but instead just revert the patch.

I hope expectations from my side here are clear.

Thanks, Daniel

Quoted 186 lines

Regards,
Sonika

-----Original Message-----
From: Daniel Vetter [mailto:[email protected]] On Behalf Of Daniel Vetter
Sent: Thursday, December 10, 2015 2:00 PM
To: Jindal, Sonika <[email protected]>
Cc: [email protected]
Subject: Re: [Intel-gfx] [PATCH] drm/i915: Add hot_plug hook for hdmi encoder

Quoted 44 lines · Sonika Jindal

From: Shashank Sharma <[email protected]>

This patch adds a separate probe function for HDMI EDID read over DDC channel. This function has been registered as a .hot_plug handler for HDMI encoder.

The current implementation of hdmi_detect() function re-sets the cached HDMI edid (in connector->detect_edid) in every detect call.This function gets called many times, sometimes directly from userspace probes, forcing drivers to read EDID every detect function call.This causes several problems like:

1. Race conditions in multiple hot_plug / unplug cases, between interrupts bottom halves and userspace detections.
2. Many Un-necessary EDID reads for single hotplug/unplug 3. HDMI complaince failures which expects only one EDID read per hotplug

This function will be serving the purpose of really reading the EDID by really probing the DDC channel, and updating the cached EDID.

The plan is to:
1. i915 IRQ handler bottom half function already calls intel_encoder->hotplug() function. Adding This probe function which will read the EDID only in case of a hotplug / unplug.
2. During init_connector this probe will be called to read the edid 3. Reuse the cached EDID in hdmi_detect() function.

The "< gen7" check is there because this was tested only for >=gen7 platforms. For older platforms the hotplug/reading edid path remains same.

v2: Calling set_edid instead of hdmi_probe during init.
Also, for platforms having DDI, intel_encoder for DP and HDMI is same
(taken from intel_dig_port), so for DP also, hot_plug function gets
called which is not intended here. So, check for HDMI in
intel_hdmi_probe Rely on HPD for updating edid only for platforms gen > 8 and also for VLV.

v3: Dropping the gen < 8 || !VLV check. Now all platforms should rely on hotplug or init for updating the edid.(Daniel) Also, calling hdmi_probe in init instead of set_edid

v4: Renaming intel_hdmi_probe to intel_hdmi_hot_plug.
Also calling this hotplug handler from intel_hpd_init to take care of init resume scenarios.

Quoted 6 lines

v5: Moved the call to encoder hotplug during init to separate patch(Daniel)
v6: Rebased and maintaining authorship.

Iirc we had to revert an earlier version of this. Commit message should cite that revert and explain why this version is better.

Also after a few days please start a new thread when resubmitting a patch series. The idea behind in-reply-to is to keep the discussion together while it's still ongoing. That's not the case when the discussion has died down since weeks.

Also, your patch to take hpd sense into account for hdmi detection has causes a regression. Your on the cc. Handling that regression needs to be absolute top priority (otherwise I need to revert that patch too).

Thanks, Daniel

Quoted 106 lines
---
 drivers/gpu/drm/i915/intel_hdmi.c | 58 
++++++++++++++++++++++++++++++---------
 1 file changed, 45 insertions(+), 13 deletions(-)
drivers/gpu/drm/i915/intel_hdmi.c +44 -13
index bdd462e..16dd2a7 100644
@@ -1381,18 +1381,16 @@ intel_hdmi_set_edid(struct drm_connector *connector, bool force)
 	return connected;
 }
 
-static enum drm_connector_status
-intel_hdmi_detect(struct drm_connector *connector, bool force)
+void intel_hdmi_hot_plug(struct intel_encoder *intel_encoder)
 {
-	enum drm_connector_status status;
-	struct intel_hdmi *intel_hdmi = intel_attached_hdmi(connector);
-	struct drm_i915_private *dev_priv = to_i915(connector->dev);
+	struct intel_hdmi *intel_hdmi =
+			enc_to_intel_hdmi(&intel_encoder->base);
+	struct intel_connector *intel_connector =
+				intel_hdmi->attached_connector;
+	struct drm_i915_private *dev_priv = 
+to_i915(intel_encoder->base.dev);
 	bool live_status = false;
 	unsigned int retry = 3;
 
-	DRM_DEBUG_KMS("[CONNECTOR:%d:%s]\n",
-		      connector->base.id, connector->name);
-
 	intel_display_power_get(dev_priv, POWER_DOMAIN_GMBUS);
 
 	while (!live_status && --retry) {
@@ -1400,21 +1398,53 @@ intel_hdmi_detect(struct drm_connector *connector, bool force)
 				hdmi_to_dig_port(intel_hdmi));
 		mdelay(10);
 	}
+	intel_display_power_put(dev_priv, POWER_DOMAIN_GMBUS);
 
 	if (!live_status)
 		DRM_DEBUG_KMS("Live status not up!");
 
-	intel_hdmi_unset_edid(connector);
+	/*
+	 * We are here, means there is a hotplug or a force
+	 * detection. Clear the cached EDID and probe the
+	 * DDC bus to check the current status of HDMI.
+	 */
+	intel_hdmi_unset_edid(&intel_connector->base);
+	if (intel_hdmi_set_edid(&intel_connector->base, live_status))
+		DRM_DEBUG_DRIVER("DDC probe: got EDID\n");
+	else
+		DRM_DEBUG_DRIVER("DDC probe: no EDID\n"); }
+
+static enum drm_connector_status
+intel_hdmi_detect(struct drm_connector *connector, bool force) {
+	enum drm_connector_status status;
+	struct intel_connector *intel_connector =
+				to_intel_connector(connector);
 
-	if (intel_hdmi_set_edid(connector, live_status)) {
+	DRM_DEBUG_KMS("[CONNECTOR:%d:%s]\n",
+		      connector->base.id, connector->name);
+
+	/*
+	 * There are many userspace calls which probe EDID from
+	 * detect path. In case of multiple hotplug/unplug, these
+	 * can cause race conditions while probing EDID. Also its
+	 * waste of CPU cycles to read the EDID again and again
+	 * unless there is a real hotplug.
+	 * So, rely on hotplugs and init to read edid.
+	 * Check connector status based on availability of cached EDID.
+	 */
+
+	if (intel_connector->detect_edid) {
 		struct intel_hdmi *intel_hdmi = intel_attached_hdmi(connector);
 
 		hdmi_to_dig_port(intel_hdmi)->base.type = INTEL_OUTPUT_HDMI;
 		status = connector_status_connected;
-	} else
+		DRM_DEBUG_DRIVER("hdmi status = connected\n");
+	} else {
 		status = connector_status_disconnected;
-
-	intel_display_power_put(dev_priv, POWER_DOMAIN_GMBUS);
+		DRM_DEBUG_DRIVER("hdmi status = disconnected\n");
+	}
 
 	return status;
 }
@@ -2131,6 +2161,8 @@ void intel_hdmi_init_connector(struct 
 intel_digital_port *intel_dig_port,
 
 	intel_hdmi_add_properties(intel_hdmi, connector);
 
+	intel_encoder->hot_plug = intel_hdmi_hot_plug;
+
 	intel_connector_attach_encoder(intel_connector, intel_encoder);
 	drm_connector_register(connector);

intel_hdmi->attached_connector = intel_connector;
--
1.9.1

Quoted 4 lines

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

--
Daniel Vetter
Software Engineer, Intel Corporation
http://blog.ffwll.ch

Signature · 7 lines
Daniel Vetter
Software Engineer, Intel Corporation
http://blog.ffwll.ch
_______________________________________________
Intel-gfx mailing list
[email protected]
http://lists.freedesktop.org/mailman/listinfo/intel-gfx
#24 lore raw

Re: [PATCH] drm/i915: Add hot_plug hook for hdmi encoder

By "we" I meant me :)
Also, since it isn't possible to check all the platforms for live status, it is safer to put a gen check to platforms which we are sure. Anyways, lets see if increasing the delay helps there, then you can take a call.

Regards,
Sonika

-----Original Message-----
From: Daniel Vetter [mailto:[email protected]] On Behalf Of Daniel Vetter
Sent: Thursday, December 10, 2015 2:24 PM
To: Jindal, Sonika <[email protected]>
Cc: Daniel Vetter <[email protected]>; [email protected]
Subject: Re: [Intel-gfx] [PATCH] drm/i915: Add hot_plug hook for hdmi encoder

Quoted 8 lines · Sonika

Sure, I will start a new thread for this.

Regarding the regression, since it is a sandybridge system, I am not sure if I can debug it anyhow because of the unavailability of the system. For the live status check, we earlier suggested to use platforms gen8 and above.
But for wider testing, we removed the check.
If required we can add a check there.

It's your patch, you need to own this. And until we've confirmed that it's a source issue I won't add a genAnything check but instead just revert the patch.

I hope expectations from my side here are clear.

Thanks, Daniel

Quoted 179 lines

Regards,
Sonika

-----Original Message-----
From: Daniel Vetter [mailto:[email protected]] On Behalf Of Daniel Vetter
Sent: Thursday, December 10, 2015 2:00 PM
To: Jindal, Sonika <[email protected]>
Cc: [email protected]
Subject: Re: [Intel-gfx] [PATCH] drm/i915: Add hot_plug hook for hdmi encoder

Quoted 44 lines · Sonika Jindal

From: Shashank Sharma <[email protected]>

This patch adds a separate probe function for HDMI EDID read over DDC channel. This function has been registered as a .hot_plug handler for HDMI encoder.

The current implementation of hdmi_detect() function re-sets the cached HDMI edid (in connector->detect_edid) in every detect call.This function gets called many times, sometimes directly from userspace probes, forcing drivers to read EDID every detect function call.This causes several problems like:

1. Race conditions in multiple hot_plug / unplug cases, between interrupts bottom halves and userspace detections.
2. Many Un-necessary EDID reads for single hotplug/unplug 3. HDMI complaince failures which expects only one EDID read per hotplug

This function will be serving the purpose of really reading the EDID by really probing the DDC channel, and updating the cached EDID.

The plan is to:
1. i915 IRQ handler bottom half function already calls intel_encoder->hotplug() function. Adding This probe function which will read the EDID only in case of a hotplug / unplug.
2. During init_connector this probe will be called to read the edid 3. Reuse the cached EDID in hdmi_detect() function.

The "< gen7" check is there because this was tested only for >=gen7 platforms. For older platforms the hotplug/reading edid path remains same.

v2: Calling set_edid instead of hdmi_probe during init.
Also, for platforms having DDI, intel_encoder for DP and HDMI is
same (taken from intel_dig_port), so for DP also, hot_plug function
gets called which is not intended here. So, check for HDMI in intel_hdmi_probe Rely on HPD for updating edid only for platforms gen > 8 and also for VLV.

v3: Dropping the gen < 8 || !VLV check. Now all platforms should rely on hotplug or init for updating the edid.(Daniel) Also, calling hdmi_probe in init instead of set_edid

v4: Renaming intel_hdmi_probe to intel_hdmi_hot_plug. Also calling this hotplug handler from intel_hpd_init to take care of init resume scenarios.

Quoted 6 lines

v5: Moved the call to encoder hotplug during init to separate patch(Daniel)
v6: Rebased and maintaining authorship.

Iirc we had to revert an earlier version of this. Commit message should cite that revert and explain why this version is better.

Also after a few days please start a new thread when resubmitting a patch series. The idea behind in-reply-to is to keep the discussion together while it's still ongoing. That's not the case when the discussion has died down since weeks.

Also, your patch to take hpd sense into account for hdmi detection has causes a regression. Your on the cc. Handling that regression needs to be absolute top priority (otherwise I need to revert that patch too).

Thanks, Daniel

Quoted 106 lines
---
 drivers/gpu/drm/i915/intel_hdmi.c | 58
++++++++++++++++++++++++++++++---------
 1 file changed, 45 insertions(+), 13 deletions(-)
drivers/gpu/drm/i915/intel_hdmi.c +44 -13
index bdd462e..16dd2a7 100644
@@ -1381,18 +1381,16 @@ intel_hdmi_set_edid(struct drm_connector *connector, bool force)
 	return connected;
 }
 
-static enum drm_connector_status
-intel_hdmi_detect(struct drm_connector *connector, bool force)
+void intel_hdmi_hot_plug(struct intel_encoder *intel_encoder)
 {
-	enum drm_connector_status status;
-	struct intel_hdmi *intel_hdmi = intel_attached_hdmi(connector);
-	struct drm_i915_private *dev_priv = to_i915(connector->dev);
+	struct intel_hdmi *intel_hdmi =
+			enc_to_intel_hdmi(&intel_encoder->base);
+	struct intel_connector *intel_connector =
+				intel_hdmi->attached_connector;
+	struct drm_i915_private *dev_priv = 
+to_i915(intel_encoder->base.dev);
 	bool live_status = false;
 	unsigned int retry = 3;
 
-	DRM_DEBUG_KMS("[CONNECTOR:%d:%s]\n",
-		      connector->base.id, connector->name);
-
 	intel_display_power_get(dev_priv, POWER_DOMAIN_GMBUS);
 
 	while (!live_status && --retry) {
@@ -1400,21 +1398,53 @@ intel_hdmi_detect(struct drm_connector *connector, bool force)
 				hdmi_to_dig_port(intel_hdmi));
 		mdelay(10);
 	}
+	intel_display_power_put(dev_priv, POWER_DOMAIN_GMBUS);
 
 	if (!live_status)
 		DRM_DEBUG_KMS("Live status not up!");
 
-	intel_hdmi_unset_edid(connector);
+	/*
+	 * We are here, means there is a hotplug or a force
+	 * detection. Clear the cached EDID and probe the
+	 * DDC bus to check the current status of HDMI.
+	 */
+	intel_hdmi_unset_edid(&intel_connector->base);
+	if (intel_hdmi_set_edid(&intel_connector->base, live_status))
+		DRM_DEBUG_DRIVER("DDC probe: got EDID\n");
+	else
+		DRM_DEBUG_DRIVER("DDC probe: no EDID\n"); }
+
+static enum drm_connector_status
+intel_hdmi_detect(struct drm_connector *connector, bool force) {
+	enum drm_connector_status status;
+	struct intel_connector *intel_connector =
+				to_intel_connector(connector);
 
-	if (intel_hdmi_set_edid(connector, live_status)) {
+	DRM_DEBUG_KMS("[CONNECTOR:%d:%s]\n",
+		      connector->base.id, connector->name);
+
+	/*
+	 * There are many userspace calls which probe EDID from
+	 * detect path. In case of multiple hotplug/unplug, these
+	 * can cause race conditions while probing EDID. Also its
+	 * waste of CPU cycles to read the EDID again and again
+	 * unless there is a real hotplug.
+	 * So, rely on hotplugs and init to read edid.
+	 * Check connector status based on availability of cached EDID.
+	 */
+
+	if (intel_connector->detect_edid) {
 		struct intel_hdmi *intel_hdmi = intel_attached_hdmi(connector);
 
 		hdmi_to_dig_port(intel_hdmi)->base.type = INTEL_OUTPUT_HDMI;
 		status = connector_status_connected;
-	} else
+		DRM_DEBUG_DRIVER("hdmi status = connected\n");
+	} else {
 		status = connector_status_disconnected;
-
-	intel_display_power_put(dev_priv, POWER_DOMAIN_GMBUS);
+		DRM_DEBUG_DRIVER("hdmi status = disconnected\n");
+	}
 
 	return status;
 }
@@ -2131,6 +2161,8 @@ void intel_hdmi_init_connector(struct 
 intel_digital_port *intel_dig_port,
 
 	intel_hdmi_add_properties(intel_hdmi, connector);
 
+	intel_encoder->hot_plug = intel_hdmi_hot_plug;
+
 	intel_connector_attach_encoder(intel_connector, intel_encoder);
 	drm_connector_register(connector);

intel_hdmi->attached_connector = intel_connector;
--
1.9.1

Quoted 9 lines
Quoted 4 lines

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

--
Daniel Vetter
Software Engineer, Intel Corporation
http://blog.ffwll.ch

--
Daniel Vetter
Software Engineer, Intel Corporation
http://blog.ffwll.ch
_______________________________________________
Intel-gfx mailing list
[email protected]
http://lists.freedesktop.org/mailman/listinfo/intel-gfx

#25 lore raw

Re: [PATCH 1/2] drm/i915: Call encoder hotplug for init and resume cases

In reply to Sonika Jindal
Quoted 28 lines · Sonika Jindal

For all the encoders, call the hot_plug if it is registered. This is required for connected boot and resume cases to generate fake hpd resulting in reading of edid.
Removing the initial sdvo hot_plug call too so that it will be called just once from this loop.

---
 drivers/gpu/drm/i915/intel_hotplug.c |   11 +++++++++++
 drivers/gpu/drm/i915/intel_sdvo.c    |    1 -
 2 files changed, 11 insertions(+), 1 deletion(-)
drivers/gpu/drm/i915/intel_hotplug.c +1 -0
index 53c0173..eac4757 100644
@@ -458,6 +458,7 @@ void intel_hpd_init(struct drm_i915_private *dev_priv)
 {
 	struct drm_device *dev = dev_priv->dev;
 	struct drm_mode_config *mode_config = &dev->mode_config;
+	struct intel_encoder *encoder;
 	struct drm_connector *connector;
 	int i;
 
@@ -482,6 +483,16 @@ void intel_hpd_init(struct drm_i915_private *dev_priv)
 	if (dev_priv->display.hpd_irq_setup)
 		dev_priv->display.hpd_irq_setup(dev);
 	spin_unlock_irq(&dev_priv->irq_lock);

drivers/gpu/drm/i915/intel_hotplug.c void intel_hpd_init(struct drm_i915_private *dev_priv) · line 495

Quoted 10 lines
drivers/gpu/drm/i915/intel_hotplug.c +10 -0
+
+	/*
+	 * Connected boot / resume scenarios can't generate new hot plug.
+	 * So, probe it manually.
+	 */
+	list_for_each_entry(encoder, &dev->mode_config.encoder_list,
+			    base.head) {
+		if (encoder->hot_plug)
+			encoder->hot_plug(encoder);
+	}

This breaks the world on CHV

[ 3187.575198] [drm:intel_hdmi_hot_plug] Live status not up!
[ 3187.585154] =============================================
[ 3187.595010] [ INFO: possible recursive locking detected ]
[ 3187.604685] 4.3.0-rc4-bsw+ #2488 Tainted: G     U  W      
[ 3187.614366] ---------------------------------------------
[ 3187.623892] Xorg/32212 is trying to acquire lock:
[ 3187.632635]  (&power_domains->lock){+.+...}, at: [<ffffffffa036aced>] intel_display_power_get+0x38/0xcb [i915]
[ 3187.647492] 
[ 3187.647492] but task is already holding lock:
[ 3187.661054]  (&power_domains->lock){+.+...}, at: [<ffffffffa036aced>] intel_display_power_get+0x38/0xcb [i915]
[ 3187.675960] 
[ 3187.675960] other info that might help us debug this:
[ 3187.690459]  Possible unsafe locking scenario:
[ 3187.690459] 
[ 3187.704224]        CPU0
[ 3187.710485]        ----
[ 3187.716711]   lock(&power_domains->lock);
[ 3187.724718]   lock(&power_domains->lock);
[ 3187.732663] 
[ 3187.732663]  *** DEADLOCK ***
[ 3187.732663] 
[ 3187.749460]  May be due to missing lock nesting notation
[ 3187.749460] 
[ 3187.763833] 5 locks held by Xorg/32212:
[ 3187.771523]  #0:  (drm_global_mutex){+.+.+.}, at: [<ffffffffa02ad5c8>] drm_release+0x3b/0x49b [drm]
[ 3187.785216]  #1:  (&dev->mode_config.mutex){+.+.+.}, at: [<ffffffffa02c4d85>] drm_modeset_lock_all+0x54/0xcd [drm]
[ 3187.800437]  #2:  (crtc_ww_class_acquire){+.+.+.}, at: [<ffffffffa02c4d8f>] drm_modeset_lock_all+0x5e/0xcd [drm]
[ 3187.815488]  #3:  (crtc_ww_class_mutex){+.+.+.}, at: [<ffffffffa02c46b0>] drm_modeset_lock+0x75/0xfc [drm]
[ 3187.830094]  #4:  (&power_domains->lock){+.+...}, at: [<ffffffffa036aced>] intel_display_power_get+0x38/0xcb [i915]
[ 3187.845534] 
[ 3187.845534] stack backtrace:
[ 3187.857685] CPU: 2 PID: 32212 Comm: Xorg Tainted: G     U  W       4.3.0-rc4-bsw+ #2488
[ 3187.870331] Hardware name: Intel Corporation CHERRYVIEW C0 PLATFORM/Braswell CRB, BIOS BRAS.X64.B085.R00.1509110553 09/11/2015
[ 3187.886827]  0000000000000000 ffff880175eff8e0 ffffffff8128d59e ffffffff823f5ee0
[ 3187.898904]  ffff880175eff958 ffffffff810a7a08 0000000000000000 ffff880179d1c5d0
[ 3187.910954]  0000000000000004 0000000000000006 45422a91588a4c3e 0000000000000005
[ 3187.923011] Call Trace:
[ 3187.929451]  [<ffffffff8128d59e>] dump_stack+0x4e/0x79
[ 3187.938912]  [<ffffffff810a7a08>] __lock_acquire+0x7ab/0x12af
[ 3187.949027]  [<ffffffff810a8d04>] lock_acquire+0x10e/0x1a9
[ 3187.958859]  [<ffffffffa036aced>] ? intel_display_power_get+0x38/0xcb [i915]
[ 3187.970476]  [<ffffffffa036aced>] ? intel_display_power_get+0x38/0xcb [i915]
[ 3187.982011]  [<ffffffff814df4a3>] mutex_lock_nested+0x71/0x346
[ 3187.992167]  [<ffffffffa036aced>] ? intel_display_power_get+0x38/0xcb [i915]
[ 3188.003655]  [<ffffffff814e20ee>] ? _raw_spin_unlock_irqrestore+0x4b/0x60
[ 3188.014829]  [<ffffffff81364565>] ? __pm_runtime_resume+0x71/0x7e
[ 3188.025269]  [<ffffffffa036aced>] intel_display_power_get+0x38/0xcb [i915]
[ 3188.036544]  [<ffffffffa036aced>] ? intel_display_power_get+0x38/0xcb [i915]
[ 3188.047968]  [<ffffffffa03e8cc4>] intel_hdmi_set_edid+0x3f/0xd6 [i915]
[ 3188.058766]  [<ffffffffa03e8e1a>] intel_hdmi_hot_plug+0xbf/0xfb [i915]
[ 3188.069484]  [<ffffffffa03c9e78>] intel_hpd_init+0xfa/0x10b [i915]
[ 3188.079753]  [<ffffffffa036968d>] vlv_display_power_well_init+0xdb/0xe8 [i915]
[ 3188.091224]  [<ffffffffa0369763>] chv_pipe_power_well_enable+0x62/0x67 [i915]
[ 3188.102594]  [<ffffffffa036ad55>] intel_display_power_get+0xa0/0xcb [i915]
[ 3188.113657]  [<ffffffffa03bdd7a>] modeset_get_crtc_power_domains+0x11d/0x13c [i915]
[ 3188.125589]  [<ffffffffa03bdfc1>] intel_atomic_commit+0x228/0xf1b [i915]
[ 3188.136522]  [<ffffffffa02c586c>] ? drm_atomic_check_only+0x37b/0x4da [drm]
[ 3188.147676]  [<ffffffffa02c5a18>] drm_atomic_commit+0x4d/0x52 [drm]
[ 3188.158023]  [<ffffffffa032fa25>] restore_fbdev_mode+0x11e/0x286 [drm_kms_helper]
[ 3188.169734]  [<ffffffffa0331443>] drm_fb_helper_restore_fbdev_mode_unlocked+0x36/0x6d [drm_kms_helper]
[ 3188.183557]  [<ffffffffa03d270f>] intel_fbdev_restore_mode+0x22/0x7a [i915]
[ 3188.194701]  [<ffffffffa03f70b8>] i915_driver_lastclose+0xe/0x15 [i915]
[ 3188.205361]  [<ffffffffa02ad4c2>] drm_lastclose+0x3f/0x10a [drm]
[ 3188.215323]  [<ffffffffa02ad9ee>] drm_release+0x461/0x49b [drm]
[ 3188.225195]  [<ffffffff8118577c>] __fput+0x100/0x1b3
[ 3188.233956]  [<ffffffff81185865>] ____fput+0xe/0x10
[ 3188.242594]  [<ffffffff81084540>] task_work_run+0x6a/0x93
[ 3188.251816]  [<ffffffff810015d6>] prepare_exit_to_usermode+0x9e/0xaf
[ 3188.262111]  [<ffffffff810017d6>] syscall_return_slowpath+0x1ef/0x264
[ 3188.272486]  [<ffffffff81084462>] ? task_work_add+0x44/0x53
[ 3188.281885]  [<ffffffff811858e3>] ? fput+0x7c/0x83
[ 3188.290389]  [<ffffffff810a6e52>] ? trace_hardirqs_on_caller+0x16/0x196
[ 3188.300941]  [<ffffffff81000c87>] ? trace_hardirqs_on_thunk+0x17/0x19
[ 3188.311245]  [<ffffffff814e2ab1>] int_ret_from_sys_call+0x25/0x9f
Quoted 22 lines
drivers/gpu/drm/i915/intel_hotplug.c +0 -0
 }
 
 void intel_hpd_init_work(struct drm_i915_private *dev_priv)
drivers/gpu/drm/i915/intel_sdvo.c +0 -1
index 05521b5..55859e9 100644
@@ -2433,7 +2433,6 @@ intel_sdvo_dvi_init(struct intel_sdvo *intel_sdvo, int device)
 		 * Ensure that they get re-enabled when an interrupt happens.
 		 */
 		intel_encoder->hot_plug = intel_sdvo_enable_hotplug;
-		intel_sdvo_enable_hotplug(intel_encoder);
 	} else {
 		intel_connector->polled = DRM_CONNECTOR_POLL_CONNECT | DRM_CONNECTOR_POLL_DISCONNECT;
 	}
Signature · 6 lines
1.7.10.4

_______________________________________________
Intel-gfx mailing list
[email protected]
http://lists.freedesktop.org/mailman/listinfo/intel-gfx
Signature · 6 lines
Ville Syrjälä
Intel OTC
_______________________________________________
Intel-gfx mailing list
[email protected]
http://lists.freedesktop.org/mailman/listinfo/intel-gfx
maintainer
#26 lore raw

Re: [PATCH 1/2] drm/i915: Call encoder hotplug for init and resume cases

Quoted 34 lines · Ville Syrjälä
Quoted 33 lines · Sonika Jindal

For all the encoders, call the hot_plug if it is registered. This is required for connected boot and resume cases to generate fake hpd resulting in reading of edid.
Removing the initial sdvo hot_plug call too so that it will be called just once from this loop.

---
 drivers/gpu/drm/i915/intel_hotplug.c |   11 +++++++++++
 drivers/gpu/drm/i915/intel_sdvo.c    |    1 -
 2 files changed, 11 insertions(+), 1 deletion(-)
drivers/gpu/drm/i915/intel_hotplug.c +6 -0
index 53c0173..eac4757 100644
@@ -458,6 +458,7 @@ void intel_hpd_init(struct drm_i915_private *dev_priv)
 {
 	struct drm_device *dev = dev_priv->dev;
 	struct drm_mode_config *mode_config = &dev->mode_config;
+	struct intel_encoder *encoder;
 	struct drm_connector *connector;
 	int i;
 
@@ -482,6 +483,16 @@ void intel_hpd_init(struct drm_i915_private *dev_priv)
 	if (dev_priv->display.hpd_irq_setup)
 		dev_priv->display.hpd_irq_setup(dev);
 	spin_unlock_irq(&dev_priv->irq_lock);
+
+	/*
+	 * Connected boot / resume scenarios can't generate new hot plug.
+	 * So, probe it manually.
+	 */
Quoted 8 lines

drivers/gpu/drm/i915/intel_hotplug.c void intel_hpd_init(struct drm_i915_private *dev_priv) · line 495

Quoted 5 lines
drivers/gpu/drm/i915/intel_hotplug.c +5 -0
+	list_for_each_entry(encoder, &dev->mode_config.encoder_list,
+			    base.head) {
+		if (encoder->hot_plug)
+			encoder->hot_plug(encoder);
+	}

This breaks the world on CHV

Also patch 2/2 breaks: https://bugs.freedesktop.org/show_bug.cgi?id=88081

BR,
Jani.

Quoted 107 lines
[ 3187.575198] [drm:intel_hdmi_hot_plug] Live status not up!
[ 3187.585154] =============================================
[ 3187.595010] [ INFO: possible recursive locking detected ]
[ 3187.604685] 4.3.0-rc4-bsw+ #2488 Tainted: G     U  W      
[ 3187.614366] ---------------------------------------------
[ 3187.623892] Xorg/32212 is trying to acquire lock:
[ 3187.632635]  (&power_domains->lock){+.+...}, at: [<ffffffffa036aced>] intel_display_power_get+0x38/0xcb [i915]
[ 3187.647492] 
[ 3187.647492] but task is already holding lock:
[ 3187.661054]  (&power_domains->lock){+.+...}, at: [<ffffffffa036aced>] intel_display_power_get+0x38/0xcb [i915]
[ 3187.675960] 
[ 3187.675960] other info that might help us debug this:
[ 3187.690459]  Possible unsafe locking scenario:
[ 3187.690459] 
[ 3187.704224]        CPU0
[ 3187.710485]        ----
[ 3187.716711]   lock(&power_domains->lock);
[ 3187.724718]   lock(&power_domains->lock);
[ 3187.732663] 
[ 3187.732663]  *** DEADLOCK ***
[ 3187.732663] 
[ 3187.749460]  May be due to missing lock nesting notation
[ 3187.749460] 
[ 3187.763833] 5 locks held by Xorg/32212:
[ 3187.771523]  #0:  (drm_global_mutex){+.+.+.}, at: [<ffffffffa02ad5c8>] drm_release+0x3b/0x49b [drm]
[ 3187.785216]  #1:  (&dev->mode_config.mutex){+.+.+.}, at: [<ffffffffa02c4d85>] drm_modeset_lock_all+0x54/0xcd [drm]
[ 3187.800437]  #2:  (crtc_ww_class_acquire){+.+.+.}, at: [<ffffffffa02c4d8f>] drm_modeset_lock_all+0x5e/0xcd [drm]
[ 3187.815488]  #3:  (crtc_ww_class_mutex){+.+.+.}, at: [<ffffffffa02c46b0>] drm_modeset_lock+0x75/0xfc [drm]
[ 3187.830094]  #4:  (&power_domains->lock){+.+...}, at: [<ffffffffa036aced>] intel_display_power_get+0x38/0xcb [i915]
[ 3187.845534] 
[ 3187.845534] stack backtrace:
[ 3187.857685] CPU: 2 PID: 32212 Comm: Xorg Tainted: G     U  W       4.3.0-rc4-bsw+ #2488
[ 3187.870331] Hardware name: Intel Corporation CHERRYVIEW C0 PLATFORM/Braswell CRB, BIOS BRAS.X64.B085.R00.1509110553 09/11/2015
[ 3187.886827]  0000000000000000 ffff880175eff8e0 ffffffff8128d59e ffffffff823f5ee0
[ 3187.898904]  ffff880175eff958 ffffffff810a7a08 0000000000000000 ffff880179d1c5d0
[ 3187.910954]  0000000000000004 0000000000000006 45422a91588a4c3e 0000000000000005
[ 3187.923011] Call Trace:
[ 3187.929451]  [<ffffffff8128d59e>] dump_stack+0x4e/0x79
[ 3187.938912]  [<ffffffff810a7a08>] __lock_acquire+0x7ab/0x12af
[ 3187.949027]  [<ffffffff810a8d04>] lock_acquire+0x10e/0x1a9
[ 3187.958859]  [<ffffffffa036aced>] ? intel_display_power_get+0x38/0xcb [i915]
[ 3187.970476]  [<ffffffffa036aced>] ? intel_display_power_get+0x38/0xcb [i915]
[ 3187.982011]  [<ffffffff814df4a3>] mutex_lock_nested+0x71/0x346
[ 3187.992167]  [<ffffffffa036aced>] ? intel_display_power_get+0x38/0xcb [i915]
[ 3188.003655]  [<ffffffff814e20ee>] ? _raw_spin_unlock_irqrestore+0x4b/0x60
[ 3188.014829]  [<ffffffff81364565>] ? __pm_runtime_resume+0x71/0x7e
[ 3188.025269]  [<ffffffffa036aced>] intel_display_power_get+0x38/0xcb [i915]
[ 3188.036544]  [<ffffffffa036aced>] ? intel_display_power_get+0x38/0xcb [i915]
[ 3188.047968]  [<ffffffffa03e8cc4>] intel_hdmi_set_edid+0x3f/0xd6 [i915]
[ 3188.058766]  [<ffffffffa03e8e1a>] intel_hdmi_hot_plug+0xbf/0xfb [i915]
[ 3188.069484]  [<ffffffffa03c9e78>] intel_hpd_init+0xfa/0x10b [i915]
[ 3188.079753]  [<ffffffffa036968d>] vlv_display_power_well_init+0xdb/0xe8 [i915]
[ 3188.091224]  [<ffffffffa0369763>] chv_pipe_power_well_enable+0x62/0x67 [i915]
[ 3188.102594]  [<ffffffffa036ad55>] intel_display_power_get+0xa0/0xcb [i915]
[ 3188.113657]  [<ffffffffa03bdd7a>] modeset_get_crtc_power_domains+0x11d/0x13c [i915]
[ 3188.125589]  [<ffffffffa03bdfc1>] intel_atomic_commit+0x228/0xf1b [i915]
[ 3188.136522]  [<ffffffffa02c586c>] ? drm_atomic_check_only+0x37b/0x4da [drm]
[ 3188.147676]  [<ffffffffa02c5a18>] drm_atomic_commit+0x4d/0x52 [drm]
[ 3188.158023]  [<ffffffffa032fa25>] restore_fbdev_mode+0x11e/0x286 [drm_kms_helper]
[ 3188.169734]  [<ffffffffa0331443>] drm_fb_helper_restore_fbdev_mode_unlocked+0x36/0x6d [drm_kms_helper]
[ 3188.183557]  [<ffffffffa03d270f>] intel_fbdev_restore_mode+0x22/0x7a [i915]
[ 3188.194701]  [<ffffffffa03f70b8>] i915_driver_lastclose+0xe/0x15 [i915]
[ 3188.205361]  [<ffffffffa02ad4c2>] drm_lastclose+0x3f/0x10a [drm]
[ 3188.215323]  [<ffffffffa02ad9ee>] drm_release+0x461/0x49b [drm]
[ 3188.225195]  [<ffffffff8118577c>] __fput+0x100/0x1b3
[ 3188.233956]  [<ffffffff81185865>] ____fput+0xe/0x10
[ 3188.242594]  [<ffffffff81084540>] task_work_run+0x6a/0x93
[ 3188.251816]  [<ffffffff810015d6>] prepare_exit_to_usermode+0x9e/0xaf
[ 3188.262111]  [<ffffffff810017d6>] syscall_return_slowpath+0x1ef/0x264
[ 3188.272486]  [<ffffffff81084462>] ? task_work_add+0x44/0x53
[ 3188.281885]  [<ffffffff811858e3>] ? fput+0x7c/0x83
[ 3188.290389]  [<ffffffff810a6e52>] ? trace_hardirqs_on_caller+0x16/0x196
[ 3188.300941]  [<ffffffff81000c87>] ? trace_hardirqs_on_thunk+0x17/0x19
[ 3188.311245]  [<ffffffff814e2ab1>] int_ret_from_sys_call+0x25/0x9f
Quoted 22 lines
drivers/gpu/drm/i915/intel_hotplug.c +0 -0
 }
 
 void intel_hpd_init_work(struct drm_i915_private *dev_priv)
drivers/gpu/drm/i915/intel_sdvo.c +0 -1
index 05521b5..55859e9 100644
@@ -2433,7 +2433,6 @@ intel_sdvo_dvi_init(struct intel_sdvo *intel_sdvo, int device)
 		 * Ensure that they get re-enabled when an interrupt happens.
 		 */
 		intel_encoder->hot_plug = intel_sdvo_enable_hotplug;
-		intel_sdvo_enable_hotplug(intel_encoder);
 	} else {
 		intel_connector->polled = DRM_CONNECTOR_POLL_CONNECT | DRM_CONNECTOR_POLL_DISCONNECT;
 	}
Signature · 6 lines
1.7.10.4

_______________________________________________
Intel-gfx mailing list
[email protected]
http://lists.freedesktop.org/mailman/listinfo/intel-gfx
Signature · 6 lines
Ville Syrjälä
Intel OTC
_______________________________________________
Intel-gfx mailing list
[email protected]
http://lists.freedesktop.org/mailman/listinfo/intel-gfx
Signature · 5 lines
Jani Nikula, Intel Open Source Technology Center
_______________________________________________
Intel-gfx mailing list
[email protected]
http://lists.freedesktop.org/mailman/listinfo/intel-gfx
#27 lore raw

Re: [PATCH 1/2] drm/i915: Call encoder hotplug for init and resume cases

Quoted 44 lines · Jani Nikula
Quoted 34 lines · Ville Syrjälä
Quoted 33 lines · Sonika Jindal

For all the encoders, call the hot_plug if it is registered. This is required for connected boot and resume cases to generate fake hpd resulting in reading of edid.
Removing the initial sdvo hot_plug call too so that it will be called just once from this loop.

---
 drivers/gpu/drm/i915/intel_hotplug.c |   11 +++++++++++
 drivers/gpu/drm/i915/intel_sdvo.c    |    1 -
 2 files changed, 11 insertions(+), 1 deletion(-)
drivers/gpu/drm/i915/intel_hotplug.c +6 -0
index 53c0173..eac4757 100644
@@ -458,6 +458,7 @@ void intel_hpd_init(struct drm_i915_private *dev_priv)
 {
 	struct drm_device *dev = dev_priv->dev;
 	struct drm_mode_config *mode_config = &dev->mode_config;
+	struct intel_encoder *encoder;
 	struct drm_connector *connector;
 	int i;
 
@@ -482,6 +483,16 @@ void intel_hpd_init(struct drm_i915_private *dev_priv)
 	if (dev_priv->display.hpd_irq_setup)
 		dev_priv->display.hpd_irq_setup(dev);
 	spin_unlock_irq(&dev_priv->irq_lock);
+
+	/*
+	 * Connected boot / resume scenarios can't generate new hot plug.
+	 * So, probe it manually.
+	 */
Quoted 8 lines

drivers/gpu/drm/i915/intel_hotplug.c void intel_hpd_init(struct drm_i915_private *dev_priv) · line 495

Quoted 5 lines
drivers/gpu/drm/i915/intel_hotplug.c +5 -0
+	list_for_each_entry(encoder, &dev->mode_config.encoder_list,
+			    base.head) {
+		if (encoder->hot_plug)
+			encoder->hot_plug(encoder);
+	}

This breaks the world on CHV

Quoted 1 line

Also patch 2/2 breaks: https://bugs.freedesktop.org/show_bug.cgi?id=88081

Both reverted, thanks for the reveport.
-Daniel

Quoted 119 lines

BR,
Jani.

Quoted 107 lines
[ 3187.575198] [drm:intel_hdmi_hot_plug] Live status not up!
[ 3187.585154] =============================================
[ 3187.595010] [ INFO: possible recursive locking detected ]
[ 3187.604685] 4.3.0-rc4-bsw+ #2488 Tainted: G     U  W      
[ 3187.614366] ---------------------------------------------
[ 3187.623892] Xorg/32212 is trying to acquire lock:
[ 3187.632635]  (&power_domains->lock){+.+...}, at: [<ffffffffa036aced>] intel_display_power_get+0x38/0xcb [i915]
[ 3187.647492] 
[ 3187.647492] but task is already holding lock:
[ 3187.661054]  (&power_domains->lock){+.+...}, at: [<ffffffffa036aced>] intel_display_power_get+0x38/0xcb [i915]
[ 3187.675960] 
[ 3187.675960] other info that might help us debug this:
[ 3187.690459]  Possible unsafe locking scenario:
[ 3187.690459] 
[ 3187.704224]        CPU0
[ 3187.710485]        ----
[ 3187.716711]   lock(&power_domains->lock);
[ 3187.724718]   lock(&power_domains->lock);
[ 3187.732663] 
[ 3187.732663]  *** DEADLOCK ***
[ 3187.732663] 
[ 3187.749460]  May be due to missing lock nesting notation
[ 3187.749460] 
[ 3187.763833] 5 locks held by Xorg/32212:
[ 3187.771523]  #0:  (drm_global_mutex){+.+.+.}, at: [<ffffffffa02ad5c8>] drm_release+0x3b/0x49b [drm]
[ 3187.785216]  #1:  (&dev->mode_config.mutex){+.+.+.}, at: [<ffffffffa02c4d85>] drm_modeset_lock_all+0x54/0xcd [drm]
[ 3187.800437]  #2:  (crtc_ww_class_acquire){+.+.+.}, at: [<ffffffffa02c4d8f>] drm_modeset_lock_all+0x5e/0xcd [drm]
[ 3187.815488]  #3:  (crtc_ww_class_mutex){+.+.+.}, at: [<ffffffffa02c46b0>] drm_modeset_lock+0x75/0xfc [drm]
[ 3187.830094]  #4:  (&power_domains->lock){+.+...}, at: [<ffffffffa036aced>] intel_display_power_get+0x38/0xcb [i915]
[ 3187.845534] 
[ 3187.845534] stack backtrace:
[ 3187.857685] CPU: 2 PID: 32212 Comm: Xorg Tainted: G     U  W       4.3.0-rc4-bsw+ #2488
[ 3187.870331] Hardware name: Intel Corporation CHERRYVIEW C0 PLATFORM/Braswell CRB, BIOS BRAS.X64.B085.R00.1509110553 09/11/2015
[ 3187.886827]  0000000000000000 ffff880175eff8e0 ffffffff8128d59e ffffffff823f5ee0
[ 3187.898904]  ffff880175eff958 ffffffff810a7a08 0000000000000000 ffff880179d1c5d0
[ 3187.910954]  0000000000000004 0000000000000006 45422a91588a4c3e 0000000000000005
[ 3187.923011] Call Trace:
[ 3187.929451]  [<ffffffff8128d59e>] dump_stack+0x4e/0x79
[ 3187.938912]  [<ffffffff810a7a08>] __lock_acquire+0x7ab/0x12af
[ 3187.949027]  [<ffffffff810a8d04>] lock_acquire+0x10e/0x1a9
[ 3187.958859]  [<ffffffffa036aced>] ? intel_display_power_get+0x38/0xcb [i915]
[ 3187.970476]  [<ffffffffa036aced>] ? intel_display_power_get+0x38/0xcb [i915]
[ 3187.982011]  [<ffffffff814df4a3>] mutex_lock_nested+0x71/0x346
[ 3187.992167]  [<ffffffffa036aced>] ? intel_display_power_get+0x38/0xcb [i915]
[ 3188.003655]  [<ffffffff814e20ee>] ? _raw_spin_unlock_irqrestore+0x4b/0x60
[ 3188.014829]  [<ffffffff81364565>] ? __pm_runtime_resume+0x71/0x7e
[ 3188.025269]  [<ffffffffa036aced>] intel_display_power_get+0x38/0xcb [i915]
[ 3188.036544]  [<ffffffffa036aced>] ? intel_display_power_get+0x38/0xcb [i915]
[ 3188.047968]  [<ffffffffa03e8cc4>] intel_hdmi_set_edid+0x3f/0xd6 [i915]
[ 3188.058766]  [<ffffffffa03e8e1a>] intel_hdmi_hot_plug+0xbf/0xfb [i915]
[ 3188.069484]  [<ffffffffa03c9e78>] intel_hpd_init+0xfa/0x10b [i915]
[ 3188.079753]  [<ffffffffa036968d>] vlv_display_power_well_init+0xdb/0xe8 [i915]
[ 3188.091224]  [<ffffffffa0369763>] chv_pipe_power_well_enable+0x62/0x67 [i915]
[ 3188.102594]  [<ffffffffa036ad55>] intel_display_power_get+0xa0/0xcb [i915]
[ 3188.113657]  [<ffffffffa03bdd7a>] modeset_get_crtc_power_domains+0x11d/0x13c [i915]
[ 3188.125589]  [<ffffffffa03bdfc1>] intel_atomic_commit+0x228/0xf1b [i915]
[ 3188.136522]  [<ffffffffa02c586c>] ? drm_atomic_check_only+0x37b/0x4da [drm]
[ 3188.147676]  [<ffffffffa02c5a18>] drm_atomic_commit+0x4d/0x52 [drm]
[ 3188.158023]  [<ffffffffa032fa25>] restore_fbdev_mode+0x11e/0x286 [drm_kms_helper]
[ 3188.169734]  [<ffffffffa0331443>] drm_fb_helper_restore_fbdev_mode_unlocked+0x36/0x6d [drm_kms_helper]
[ 3188.183557]  [<ffffffffa03d270f>] intel_fbdev_restore_mode+0x22/0x7a [i915]
[ 3188.194701]  [<ffffffffa03f70b8>] i915_driver_lastclose+0xe/0x15 [i915]
[ 3188.205361]  [<ffffffffa02ad4c2>] drm_lastclose+0x3f/0x10a [drm]
[ 3188.215323]  [<ffffffffa02ad9ee>] drm_release+0x461/0x49b [drm]
[ 3188.225195]  [<ffffffff8118577c>] __fput+0x100/0x1b3
[ 3188.233956]  [<ffffffff81185865>] ____fput+0xe/0x10
[ 3188.242594]  [<ffffffff81084540>] task_work_run+0x6a/0x93
[ 3188.251816]  [<ffffffff810015d6>] prepare_exit_to_usermode+0x9e/0xaf
[ 3188.262111]  [<ffffffff810017d6>] syscall_return_slowpath+0x1ef/0x264
[ 3188.272486]  [<ffffffff81084462>] ? task_work_add+0x44/0x53
[ 3188.281885]  [<ffffffff811858e3>] ? fput+0x7c/0x83
[ 3188.290389]  [<ffffffff810a6e52>] ? trace_hardirqs_on_caller+0x16/0x196
[ 3188.300941]  [<ffffffff81000c87>] ? trace_hardirqs_on_thunk+0x17/0x19
[ 3188.311245]  [<ffffffff814e2ab1>] int_ret_from_sys_call+0x25/0x9f
Quoted 22 lines
drivers/gpu/drm/i915/intel_hotplug.c +0 -0
 }
 
 void intel_hpd_init_work(struct drm_i915_private *dev_priv)
drivers/gpu/drm/i915/intel_sdvo.c +0 -1
index 05521b5..55859e9 100644
@@ -2433,7 +2433,6 @@ intel_sdvo_dvi_init(struct intel_sdvo *intel_sdvo, int device)
 		 * Ensure that they get re-enabled when an interrupt happens.
 		 */
 		intel_encoder->hot_plug = intel_sdvo_enable_hotplug;
-		intel_sdvo_enable_hotplug(intel_encoder);
 	} else {
 		intel_connector->polled = DRM_CONNECTOR_POLL_CONNECT | DRM_CONNECTOR_POLL_DISCONNECT;
 	}
Signature · 6 lines
1.7.10.4

_______________________________________________
Intel-gfx mailing list
[email protected]
http://lists.freedesktop.org/mailman/listinfo/intel-gfx
Signature · 6 lines
Ville Syrjälä
Intel OTC
_______________________________________________
Intel-gfx mailing list
[email protected]
http://lists.freedesktop.org/mailman/listinfo/intel-gfx
Signature · 5 lines
Jani Nikula, Intel Open Source Technology Center
_______________________________________________
Intel-gfx mailing list
[email protected]
http://lists.freedesktop.org/mailman/listinfo/intel-gfx
Signature · 7 lines
Daniel Vetter
Software Engineer, Intel Corporation
http://blog.ffwll.ch
_______________________________________________
Intel-gfx mailing list
[email protected]
http://lists.freedesktop.org/mailman/listinfo/intel-gfx
#28 lore raw

Re: [PATCH 1/2] drm/i915: Call encoder hotplug for init and resume cases

Quoted 45 lines · Daniel Vetter
Quoted 44 lines · Jani Nikula
Quoted 42 lines · Ville Syrjälä
Quoted 28 lines · Sonika Jindal

For all the encoders, call the hot_plug if it is registered. This is required for connected boot and resume cases to generate fake hpd resulting in reading of edid.
Removing the initial sdvo hot_plug call too so that it will be called just once from this loop.

---
  drivers/gpu/drm/i915/intel_hotplug.c |   11 +++++++++++
  drivers/gpu/drm/i915/intel_sdvo.c    |    1 -
  2 files changed, 11 insertions(+), 1 deletion(-)
drivers/gpu/drm/i915/intel_hotplug.c +1 -0
index 53c0173..eac4757 100644
@@ -458,6 +458,7 @@ void intel_hpd_init(struct drm_i915_private *dev_priv)
  {
  	struct drm_device *dev = dev_priv->dev;
  	struct drm_mode_config *mode_config = &dev->mode_config;
+	struct intel_encoder *encoder;
  	struct drm_connector *connector;
  	int i;
 
@@ -482,6 +483,16 @@ void intel_hpd_init(struct drm_i915_private *dev_priv)
  	if (dev_priv->display.hpd_irq_setup)
  		dev_priv->display.hpd_irq_setup(dev);
  	spin_unlock_irq(&dev_priv->irq_lock);

drivers/gpu/drm/i915/intel_hotplug.c void intel_hpd_init(struct drm_i915_private *dev_priv) · line 495

Quoted 10 lines
drivers/gpu/drm/i915/intel_hotplug.c +10 -0
+
+	/*
+	 * Connected boot / resume scenarios can't generate new hot plug.
+	 * So, probe it manually.
+	 */
+	list_for_each_entry(encoder, &dev->mode_config.encoder_list,
+			    base.head) {
+		if (encoder->hot_plug)
+			encoder->hot_plug(encoder);
+	}

This breaks the world on CHV

Quoted 2 lines
Quoted 1 line

Also patch 2/2 breaks: https://bugs.freedesktop.org/show_bug.cgi?id=88081

Hmm, will check why live status is not up for chv..

Regards,
Sonika

Quoted 123 lines

Both reverted, thanks for the reveport.
-Daniel

Quoted 119 lines

BR,
Jani.

Quoted 95 lines
[ 3187.575198] [drm:intel_hdmi_hot_plug] Live status not up!
[ 3187.585154] =============================================
[ 3187.595010] [ INFO: possible recursive locking detected ]
[ 3187.604685] 4.3.0-rc4-bsw+ #2488 Tainted: G     U  W
[ 3187.614366] ---------------------------------------------
[ 3187.623892] Xorg/32212 is trying to acquire lock:
[ 3187.632635]  (&power_domains->lock){+.+...}, at: [<ffffffffa036aced>] intel_display_power_get+0x38/0xcb [i915]
[ 3187.647492]
[ 3187.647492] but task is already holding lock:
[ 3187.661054]  (&power_domains->lock){+.+...}, at: [<ffffffffa036aced>] intel_display_power_get+0x38/0xcb [i915]
[ 3187.675960]
[ 3187.675960] other info that might help us debug this:
[ 3187.690459]  Possible unsafe locking scenario:
[ 3187.690459]
[ 3187.704224]        CPU0
[ 3187.710485]        ----
[ 3187.716711]   lock(&power_domains->lock);
[ 3187.724718]   lock(&power_domains->lock);
[ 3187.732663]
[ 3187.732663]  *** DEADLOCK ***
[ 3187.732663]
[ 3187.749460]  May be due to missing lock nesting notation
[ 3187.749460]
[ 3187.763833] 5 locks held by Xorg/32212:
[ 3187.771523]  #0:  (drm_global_mutex){+.+.+.}, at: [<ffffffffa02ad5c8>] drm_release+0x3b/0x49b [drm]
[ 3187.785216]  #1:  (&dev->mode_config.mutex){+.+.+.}, at: [<ffffffffa02c4d85>] drm_modeset_lock_all+0x54/0xcd [drm]
[ 3187.800437]  #2:  (crtc_ww_class_acquire){+.+.+.}, at: [<ffffffffa02c4d8f>] drm_modeset_lock_all+0x5e/0xcd [drm]
[ 3187.815488]  #3:  (crtc_ww_class_mutex){+.+.+.}, at: [<ffffffffa02c46b0>] drm_modeset_lock+0x75/0xfc [drm]
[ 3187.830094]  #4:  (&power_domains->lock){+.+...}, at: [<ffffffffa036aced>] intel_display_power_get+0x38/0xcb [i915]
[ 3187.845534]
[ 3187.845534] stack backtrace:
[ 3187.857685] CPU: 2 PID: 32212 Comm: Xorg Tainted: G     U  W       4.3.0-rc4-bsw+ #2488
[ 3187.870331] Hardware name: Intel Corporation CHERRYVIEW C0 PLATFORM/Braswell CRB, BIOS BRAS.X64.B085.R00.1509110553 09/11/2015
[ 3187.886827]  0000000000000000 ffff880175eff8e0 ffffffff8128d59e ffffffff823f5ee0
[ 3187.898904]  ffff880175eff958 ffffffff810a7a08 0000000000000000 ffff880179d1c5d0
[ 3187.910954]  0000000000000004 0000000000000006 45422a91588a4c3e 0000000000000005
[ 3187.923011] Call Trace:
[ 3187.929451]  [<ffffffff8128d59e>] dump_stack+0x4e/0x79
[ 3187.938912]  [<ffffffff810a7a08>] __lock_acquire+0x7ab/0x12af
[ 3187.949027]  [<ffffffff810a8d04>] lock_acquire+0x10e/0x1a9
[ 3187.958859]  [<ffffffffa036aced>] ? intel_display_power_get+0x38/0xcb [i915]
[ 3187.970476]  [<ffffffffa036aced>] ? intel_display_power_get+0x38/0xcb [i915]
[ 3187.982011]  [<ffffffff814df4a3>] mutex_lock_nested+0x71/0x346
[ 3187.992167]  [<ffffffffa036aced>] ? intel_display_power_get+0x38/0xcb [i915]
[ 3188.003655]  [<ffffffff814e20ee>] ? _raw_spin_unlock_irqrestore+0x4b/0x60
[ 3188.014829]  [<ffffffff81364565>] ? __pm_runtime_resume+0x71/0x7e
[ 3188.025269]  [<ffffffffa036aced>] intel_display_power_get+0x38/0xcb [i915]
[ 3188.036544]  [<ffffffffa036aced>] ? intel_display_power_get+0x38/0xcb [i915]
[ 3188.047968]  [<ffffffffa03e8cc4>] intel_hdmi_set_edid+0x3f/0xd6 [i915]
[ 3188.058766]  [<ffffffffa03e8e1a>] intel_hdmi_hot_plug+0xbf/0xfb [i915]
[ 3188.069484]  [<ffffffffa03c9e78>] intel_hpd_init+0xfa/0x10b [i915]
[ 3188.079753]  [<ffffffffa036968d>] vlv_display_power_well_init+0xdb/0xe8 [i915]
[ 3188.091224]  [<ffffffffa0369763>] chv_pipe_power_well_enable+0x62/0x67 [i915]
[ 3188.102594]  [<ffffffffa036ad55>] intel_display_power_get+0xa0/0xcb [i915]
[ 3188.113657]  [<ffffffffa03bdd7a>] modeset_get_crtc_power_domains+0x11d/0x13c [i915]
[ 3188.125589]  [<ffffffffa03bdfc1>] intel_atomic_commit+0x228/0xf1b [i915]
[ 3188.136522]  [<ffffffffa02c586c>] ? drm_atomic_check_only+0x37b/0x4da [drm]
[ 3188.147676]  [<ffffffffa02c5a18>] drm_atomic_commit+0x4d/0x52 [drm]
[ 3188.158023]  [<ffffffffa032fa25>] restore_fbdev_mode+0x11e/0x286 [drm_kms_helper]
[ 3188.169734]  [<ffffffffa0331443>] drm_fb_helper_restore_fbdev_mode_unlocked+0x36/0x6d [drm_kms_helper]
[ 3188.183557]  [<ffffffffa03d270f>] intel_fbdev_restore_mode+0x22/0x7a [i915]
[ 3188.194701]  [<ffffffffa03f70b8>] i915_driver_lastclose+0xe/0x15 [i915]
[ 3188.205361]  [<ffffffffa02ad4c2>] drm_lastclose+0x3f/0x10a [drm]
[ 3188.215323]  [<ffffffffa02ad9ee>] drm_release+0x461/0x49b [drm]
[ 3188.225195]  [<ffffffff8118577c>] __fput+0x100/0x1b3
[ 3188.233956]  [<ffffffff81185865>] ____fput+0xe/0x10
[ 3188.242594]  [<ffffffff81084540>] task_work_run+0x6a/0x93
[ 3188.251816]  [<ffffffff810015d6>] prepare_exit_to_usermode+0x9e/0xaf
[ 3188.262111]  [<ffffffff810017d6>] syscall_return_slowpath+0x1ef/0x264
[ 3188.272486]  [<ffffffff81084462>] ? task_work_add+0x44/0x53
[ 3188.281885]  [<ffffffff811858e3>] ? fput+0x7c/0x83
[ 3188.290389]  [<ffffffff810a6e52>] ? trace_hardirqs_on_caller+0x16/0x196
[ 3188.300941]  [<ffffffff81000c87>] ? trace_hardirqs_on_thunk+0x17/0x19
[ 3188.311245]  [<ffffffff814e2ab1>] int_ret_from_sys_call+0x25/0x9f
Quoted 18 lines
drivers/gpu/drm/i915/intel_hotplug.c +0 -0
  }
 
  void intel_hpd_init_work(struct drm_i915_private *dev_priv)
drivers/gpu/drm/i915/intel_sdvo.c +0 -2
index 05521b5..55859e9 100644
@@ -2433,7 +2433,6 @@ intel_sdvo_dvi_init(struct intel_sdvo *intel_sdvo, int device)
  		 * Ensure that they get re-enabled when an interrupt happens.
  		 */
  		intel_encoder->hot_plug = intel_sdvo_enable_hotplug;
-		intel_sdvo_enable_hotplug(intel_encoder);
  	} else {
  		intel_connector->polled = DRM_CONNECTOR_POLL_CONNECT | DRM_CONNECTOR_POLL_DISCONNECT;
  	}
--

1.7.10.4

Quoted 12 lines
Quoted 4 lines

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

--
Ville Syrjälä
Intel OTC
_______________________________________________
Intel-gfx mailing list
[email protected]
http://lists.freedesktop.org/mailman/listinfo/intel-gfx

--
Jani Nikula, Intel Open Source Technology Center
_______________________________________________
Intel-gfx mailing list
[email protected]
http://lists.freedesktop.org/mailman/listinfo/intel-gfx

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

#29 lore raw

PATCH 1/2 drm/i915: Call encoder hotplug for init and resume cases

In reply to Daniel Vetter

Call the encoders, call the hot_plug if it is registered. This is required for connected boot and resume cases to generate fake hpd resulting in reading of edid.
Removing the initial sdvo hot_plug call too so that it will be called just once from this loop.

v2: Schedule a work function to call hot_plug

---
 drivers/gpu/drm/i915/i915_drv.h      |  1 +
 drivers/gpu/drm/i915/intel_hotplug.c | 26 ++++++++++++++++++++++++++
 drivers/gpu/drm/i915/intel_sdvo.c    |  1 -
 3 files changed, 27 insertions(+), 1 deletion(-)
drivers/gpu/drm/i915/i915_drv.h +1 -0
index bc865e23..4f037b9 100644
@@ -228,6 +228,7 @@ enum hpd_pin {
 
 struct i915_hotplug {
 	struct work_struct hotplug_work;
+	struct work_struct edid_work;
 
 	struct {
 		unsigned long last_jiffies;
drivers/gpu/drm/i915/intel_hotplug.c +26 -0
index b177857..72d8fe8 100644
@@ -442,6 +442,24 @@ void intel_hpd_irq_handler(struct drm_device *dev,
 		schedule_work(&dev_priv->hotplug.hotplug_work);
 }
 
+static void i915_edid_work_func(struct work_struct *work)
+{
+	struct drm_i915_private *dev_priv =
+		container_of(work, struct drm_i915_private, hotplug.edid_work);
+	struct drm_device *dev = dev_priv->dev;
+	struct drm_mode_config *mode_config = &dev->mode_config;
+	struct intel_encoder *encoder;
+
+	mutex_lock(&mode_config->mutex);
+	list_for_each_entry(encoder, &mode_config->encoder_list,
+			base.head) {
+		if (encoder->hot_plug)
+			encoder->hot_plug(encoder);
+	}
+	mutex_unlock(&mode_config->mutex);
+	drm_helper_hpd_irq_event(dev);
+}
+
 /**
  * intel_hpd_init - initializes and enables hpd support
  * @dev_priv: i915 device instance
@@ -482,12 +500,19 @@ void intel_hpd_init(struct drm_i915_private *dev_priv)
 	if (dev_priv->display.hpd_irq_setup)
 		dev_priv->display.hpd_irq_setup(dev);
 	spin_unlock_irq(&dev_priv->irq_lock);
+
+	/*
+	 * Connected boot / resume scenarios can't generate new hot plug.
+	 * So, probe it manually.
+	 */
+	schedule_work(&dev_priv->hotplug.edid_work);
 }
 
 void intel_hpd_init_work(struct drm_i915_private *dev_priv)
 {
 	INIT_WORK(&dev_priv->hotplug.hotplug_work, i915_hotplug_work_func);
 	INIT_WORK(&dev_priv->hotplug.dig_port_work, i915_digport_work_func);
+	INIT_WORK(&dev_priv->hotplug.edid_work, i915_edid_work_func);
 	INIT_DELAYED_WORK(&dev_priv->hotplug.reenable_work,
 			  intel_hpd_irq_storm_reenable_work);
 }
@@ -504,5 +529,6 @@ void intel_hpd_cancel_work(struct drm_i915_private *dev_priv)
 
 	cancel_work_sync(&dev_priv->hotplug.dig_port_work);
 	cancel_work_sync(&dev_priv->hotplug.hotplug_work);
+	cancel_work_sync(&dev_priv->hotplug.edid_work);
 	cancel_delayed_work_sync(&dev_priv->hotplug.reenable_work);
 }
drivers/gpu/drm/i915/intel_sdvo.c +0 -1
index 06679f1..4238a02 100644
@@ -2466,7 +2466,6 @@ intel_sdvo_dvi_init(struct intel_sdvo *intel_sdvo, int device)
 		 * Ensure that they get re-enabled when an interrupt happens.
 		 */
 		intel_encoder->hot_plug = intel_sdvo_enable_hotplug;
-		intel_sdvo_enable_hotplug(intel_encoder);
 	} else {
 		intel_connector->polled = DRM_CONNECTOR_POLL_CONNECT | DRM_CONNECTOR_POLL_DISCONNECT;
 	}
Signature · 6 lines
1.9.1

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

Re: [PATCH 1/2] drm/i915: Call encoder hotplug for init and resume cases

In reply to Ville Syrjälä

We were debugging this issue, and we could find the root cause:
In function:
Intel_hpd_init()

	|
encoder->hotplug()
	|
display_power_get()
	|
Intel_powe_well_enable()
	|
power_well->ops->enable()
	|
chv_pipe_power_well_enable()
	|
vlv_display_power_well_init()
	|

intel_hdp_init()

This function ends up calling intel_hpd_init, Which is causing the mutex deadlock due to recursion, as we are calling encoder->hotplug() from hpd_init().

Intel_hpd_init()

	|
encoder->hotplug()
	|

display_power_get()

There are two solutions here:
- remove hpd_init() from get_calls
- remove encoder->hotplug() function call from hpd_init() and put it in some other place. We have added this function from encoder_init() for android trees, and it works well there.

Regards
Shashank
-----Original Message-----
From: Intel-gfx [mailto:[email protected]] On Behalf Of Ville Syrjälä
Sent: Thursday, October 08, 2015 7:06 PM
To: Jindal, Sonika
Cc: [email protected]
Subject: Re: [Intel-gfx] [PATCH 1/2] drm/i915: Call encoder hotplug for init and resume cases

Quoted 29 lines · Sonika Jindal

For all the encoders, call the hot_plug if it is registered. This is required for connected boot and resume cases to generate fake hpd resulting in reading of edid.
Removing the initial sdvo hot_plug call too so that it will be called just once from this loop.

---
 drivers/gpu/drm/i915/intel_hotplug.c |   11 +++++++++++
 drivers/gpu/drm/i915/intel_sdvo.c    |    1 -
 2 files changed, 11 insertions(+), 1 deletion(-)
drivers/gpu/drm/i915/intel_hotplug.c +1 -0
index 53c0173..eac4757 100644
@@ -458,6 +458,7 @@ void intel_hpd_init(struct drm_i915_private 
 *dev_priv)  {
 	struct drm_device *dev = dev_priv->dev;
 	struct drm_mode_config *mode_config = &dev->mode_config;
+	struct intel_encoder *encoder;
 	struct drm_connector *connector;
 	int i;
 
@@ -482,6 +483,16 @@ void intel_hpd_init(struct drm_i915_private *dev_priv)
 	if (dev_priv->display.hpd_irq_setup)
 		dev_priv->display.hpd_irq_setup(dev);
 	spin_unlock_irq(&dev_priv->irq_lock);

drivers/gpu/drm/i915/intel_hotplug.c void intel_hpd_init(struct drm_i915_private *dev_priv) · line 495

Quoted 10 lines
drivers/gpu/drm/i915/intel_hotplug.c +10 -0
+
+	/*
+	 * Connected boot / resume scenarios can't generate new hot plug.
+	 * So, probe it manually.
+	 */
+	list_for_each_entry(encoder, &dev->mode_config.encoder_list,
+			    base.head) {
+		if (encoder->hot_plug)
+			encoder->hot_plug(encoder);
+	}

This breaks the world on CHV

[ 3187.575198] [drm:intel_hdmi_hot_plug] Live status not up!
[ 3187.585154] =============================================
[ 3187.595010] [ INFO: possible recursive locking detected ]
[ 3187.604685] 4.3.0-rc4-bsw+ #2488 Tainted: G     U  W      
[ 3187.614366] ---------------------------------------------
[ 3187.623892] Xorg/32212 is trying to acquire lock:
[ 3187.632635]  (&power_domains->lock){+.+...}, at: [<ffffffffa036aced>] intel_display_power_get+0x38/0xcb [i915] [ 3187.647492] [ 3187.647492] but task is already holding lock:
[ 3187.661054]  (&power_domains->lock){+.+...}, at: [<ffffffffa036aced>] intel_display_power_get+0x38/0xcb [i915] [ 3187.675960] [ 3187.675960] other info that might help us debug this:
[ 3187.690459]  Possible unsafe locking scenario:
[ 3187.690459] 
[ 3187.704224]        CPU0
[ 3187.710485]        ----
[ 3187.716711]   lock(&power_domains->lock);
[ 3187.724718]   lock(&power_domains->lock);
[ 3187.732663]
[ 3187.732663]  *** DEADLOCK ***
[ 3187.732663]
[ 3187.749460]  May be due to missing lock nesting notation [ 3187.749460] [ 3187.763833] 5 locks held by Xorg/32212:
[ 3187.771523]  #0:  (drm_global_mutex){+.+.+.}, at: [<ffffffffa02ad5c8>] drm_release+0x3b/0x49b [drm] [ 3187.785216]  #1:  (&dev->mode_config.mutex){+.+.+.}, at: [<ffffffffa02c4d85>] drm_modeset_lock_all+0x54/0xcd [drm] [ 3187.800437]  #2:  (crtc_ww_class_acquire){+.+.+.}, at: [<ffffffffa02c4d8f>] drm_modeset_lock_all+0x5e/0xcd [drm] [ 3187.815488]  #3:  (crtc_ww_class_mutex){+.+.+.}, at: [<ffffffffa02c46b0>] drm_modeset_lock+0x75/0xfc [drm] [ 3187.830094]  #4:  (&power_domains->lock){+.+...}, at: [<ffffffffa036aced>] intel_display_power_get+0x38/0xcb [i915] [ 3187.845534] [ 3187.845534] stack backtrace:
[ 3187.857685] CPU: 2 PID: 32212 Comm: Xorg Tainted: G     U  W       4.3.0-rc4-bsw+ #2488
[ 3187.870331] Hardware name: Intel Corporation CHERRYVIEW C0 PLATFORM/Braswell CRB, BIOS BRAS.X64.B085.R00.1509110553 09/11/2015 [ 3187.886827]  0000000000000000 ffff880175eff8e0 ffffffff8128d59e ffffffff823f5ee0 [ 3187.898904]  ffff880175eff958 ffffffff810a7a08 0000000000000000 ffff880179d1c5d0 [ 3187.910954]  0000000000000004 0000000000000006 45422a91588a4c3e 0000000000000005 [ 3187.923011] Call Trace:
[ 3187.929451]  [<ffffffff8128d59e>] dump_stack+0x4e/0x79 [ 3187.938912]  [<ffffffff810a7a08>] __lock_acquire+0x7ab/0x12af [ 3187.949027]  [<ffffffff810a8d04>] lock_acquire+0x10e/0x1a9 [ 3187.958859]  [<ffffffffa036aced>] ? intel_display_power_get+0x38/0xcb [i915] [ 3187.970476]  [<ffffffffa036aced>] ? intel_display_power_get+0x38/0xcb [i915] [ 3187.982011]  [<ffffffff814df4a3>] mutex_lock_nested+0x71/0x346 [ 3187.992167]  [<ffffffffa036aced>] ? intel_display_power_get+0x38/0xcb [i915] [ 3188.003655]  [<ffffffff814e20ee>] ? _raw_spin_unlock_irqrestore+0x4b/0x60
[ 3188.014829]  [<ffffffff81364565>] ? __pm_runtime_resume+0x71/0x7e [ 3188.025269]  [<ffffffffa036aced>] intel_display_power_get+0x38/0xcb [i915] [ 3188.036544]  [<ffffffffa036aced>] ? intel_display_power_get+0x38/0xcb [i915] [ 3188.047968]  [<ffffffffa03e8cc4>] intel_hdmi_set_edid+0x3f/0xd6 [i915] [ 3188.058766]  [<ffffffffa03e8e1a>] intel_hdmi_hot_plug+0xbf/0xfb [i915] [ 3188.069484]  [<ffffffffa03c9e78>] intel_hpd_init+0xfa/0x10b [i915] [ 3188.079753]  [<ffffffffa036968d>] vlv_display_power_well_init+0xdb/0xe8 [i915] [ 3188.091224]  [<ffffffffa0369763>] chv_pipe_power_well_enable+0x62/0x67 [i915] [ 3188.102594]  [<ffffffffa036ad55>] intel_display_power_get+0xa0/0xcb [i915] [ 3188.113657]  [<ffffffffa03bdd7a>] modeset_get_crtc_power_domains+0x11d/0x13c [i915] [ 3188.125589]  [<ffffffffa03bdfc1>] intel_atomic_commit+0x228/0xf1b [i915] [ 3188.136522]  [<ffffffffa02c586c>] ? drm_atomic_check_only+0x37b/0x4da [drm] [ 3188.147676]  [<ffffffffa02c5a18>] drm_atomic_commit+0x4d/0x52 [drm] [ 3188.158023]  [<ffffffffa032fa25>] restore_fbdev_mode+0x11e/0x286 [drm_kms_helper] [ 3188.169734]  [<ffffffffa0331443>] drm_fb_helper_restore_fbdev_mode_unlocked+0x36/0x6d [drm_kms_helper] [ 3188.183557]  [<ffffffffa03d270f>] intel_fbdev_restore_mode+0x22/0x7a [i915] [ 3188.194701]  [<ffffffffa03f70b8>] i915_driver_lastclose+0xe/0x15 [i915] [ 3188.205361]  [<ffffffffa02ad4c2>] drm_lastclose+0x3f/0x10a [drm] [ 3188.215323]  [<ffffffffa02ad9ee>] drm_release+0x461/0x49b [drm] [ 3188.225195]  [<ffffffff8118577c>] __fput+0x100/0x1b3 [ 3188.233956]  [<ffffffff81185865>] ____fput+0xe/0x10 [ 3188.242594]  [<ffffffff81084540>] task_work_run+0x6a/0x93 [ 3188.251816]  [<ffffffff810015d6>] prepare_exit_to_usermode+0x9e/0xaf
[ 3188.262111]  [<ffffffff810017d6>] syscall_return_slowpath+0x1ef/0x264
[ 3188.272486]  [<ffffffff81084462>] ? task_work_add+0x44/0x53 [ 3188.281885]  [<ffffffff811858e3>] ? fput+0x7c/0x83 [ 3188.290389]  [<ffffffff810a6e52>] ? trace_hardirqs_on_caller+0x16/0x196
[ 3188.300941]  [<ffffffff81000c87>] ? trace_hardirqs_on_thunk+0x17/0x19 [ 3188.311245]  [<ffffffff814e2ab1>] int_ret_from_sys_call+0x25/0x9f
Quoted 19 lines
drivers/gpu/drm/i915/intel_hotplug.c +0 -1
 }
 
 void intel_hpd_init_work(struct drm_i915_private *dev_priv) diff 
--git a/drivers/gpu/drm/i915/intel_sdvo.c 

b/drivers/gpu/drm/i915/intel_sdvo.c
index 05521b5..55859e9 100644

drivers/gpu/drm/i915/intel_sdvo.c +0 -2
@@ -2433,7 +2433,6 @@ intel_sdvo_dvi_init(struct intel_sdvo *intel_sdvo, int device)
 		 * Ensure that they get re-enabled when an interrupt happens.
 		 */
 		intel_encoder->hot_plug = intel_sdvo_enable_hotplug;
-		intel_sdvo_enable_hotplug(intel_encoder);
 	} else {
 		intel_connector->polled = DRM_CONNECTOR_POLL_CONNECT | DRM_CONNECTOR_POLL_DISCONNECT;
 	}
--

1.7.10.4

Quoted 4 lines

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

--
Ville Syrjälä
Intel OTC
_______________________________________________
Intel-gfx mailing list
[email protected]
http://lists.freedesktop.org/mailman/listinfo/intel-gfx
_______________________________________________
Intel-gfx mailing list
[email protected]
http://lists.freedesktop.org/mailman/listinfo/intel-gfx

#31 lore raw

Re: [PATCH 1/2] drm/i915: Call encoder hotplug for init and resume cases

Had some discussion with Daniel on IRC about how we can fix the deadlock.
We couldn't decide upon any solution but I can give it a try.
He suggested to have a work function for ->hot_plug which will get the power domain lock and not allow ->hot_plug if the power_domain is shut. Then when it calls hpt_init again from power_domain get, ->hot_plug gets called.
But this looks racy, we can try this out.
If there are any other suggestions, please let me know.
I will work on this after 26th Oct.

Regards,
Sonika

-----Original Message-----
From: Sharma, Shashank
Sent: Monday, October 12, 2015 5:54 PM
To: Ville Syrjälä; Vetter, Daniel
Cc: [email protected]; Mukherjee, Indranil; Jindal, Sonika
Subject: RE: [Intel-gfx] [PATCH 1/2] drm/i915: Call encoder hotplug for init and resume cases

We were debugging this issue, and we could find the root cause:
In function:
Intel_hpd_init()

	|
encoder->hotplug()
	|
display_power_get()
	|
Intel_powe_well_enable()
	|
power_well->ops->enable()
	|
chv_pipe_power_well_enable()
	|
vlv_display_power_well_init()
	|

intel_hdp_init()

This function ends up calling intel_hpd_init, Which is causing the mutex deadlock due to recursion, as we are calling encoder->hotplug() from hpd_init().

Intel_hpd_init()

	|
encoder->hotplug()
	|

display_power_get()

There are two solutions here:
- remove hpd_init() from get_calls
- remove encoder->hotplug() function call from hpd_init() and put it in some other place. We have added this function from encoder_init() for android trees, and it works well there.

Regards
Shashank
-----Original Message-----
From: Intel-gfx [mailto:[email protected]] On Behalf Of Ville Syrjälä
Sent: Thursday, October 08, 2015 7:06 PM
To: Jindal, Sonika
Cc: [email protected]
Subject: Re: [Intel-gfx] [PATCH 1/2] drm/i915: Call encoder hotplug for init and resume cases

Quoted 29 lines · Sonika Jindal

For all the encoders, call the hot_plug if it is registered. This is required for connected boot and resume cases to generate fake hpd resulting in reading of edid.
Removing the initial sdvo hot_plug call too so that it will be called just once from this loop.

---
 drivers/gpu/drm/i915/intel_hotplug.c |   11 +++++++++++
 drivers/gpu/drm/i915/intel_sdvo.c    |    1 -
 2 files changed, 11 insertions(+), 1 deletion(-)
drivers/gpu/drm/i915/intel_hotplug.c +1 -0
index 53c0173..eac4757 100644
@@ -458,6 +458,7 @@ void intel_hpd_init(struct drm_i915_private
 *dev_priv)  {
 	struct drm_device *dev = dev_priv->dev;
 	struct drm_mode_config *mode_config = &dev->mode_config;
+	struct intel_encoder *encoder;
 	struct drm_connector *connector;
 	int i;
 
@@ -482,6 +483,16 @@ void intel_hpd_init(struct drm_i915_private *dev_priv)
 	if (dev_priv->display.hpd_irq_setup)
 		dev_priv->display.hpd_irq_setup(dev);
 	spin_unlock_irq(&dev_priv->irq_lock);

drivers/gpu/drm/i915/intel_hotplug.c void intel_hpd_init(struct drm_i915_private *dev_priv) · line 495

Quoted 10 lines
drivers/gpu/drm/i915/intel_hotplug.c +10 -0
+
+	/*
+	 * Connected boot / resume scenarios can't generate new hot plug.
+	 * So, probe it manually.
+	 */
+	list_for_each_entry(encoder, &dev->mode_config.encoder_list,
+			    base.head) {
+		if (encoder->hot_plug)
+			encoder->hot_plug(encoder);
+	}

This breaks the world on CHV

[ 3187.575198] [drm:intel_hdmi_hot_plug] Live status not up!
[ 3187.585154] =============================================
[ 3187.595010] [ INFO: possible recursive locking detected ]
[ 3187.604685] 4.3.0-rc4-bsw+ #2488 Tainted: G     U  W      
[ 3187.614366] ---------------------------------------------
[ 3187.623892] Xorg/32212 is trying to acquire lock:
[ 3187.632635]  (&power_domains->lock){+.+...}, at: [<ffffffffa036aced>] intel_display_power_get+0x38/0xcb [i915] [ 3187.647492] [ 3187.647492] but task is already holding lock:
[ 3187.661054]  (&power_domains->lock){+.+...}, at: [<ffffffffa036aced>] intel_display_power_get+0x38/0xcb [i915] [ 3187.675960] [ 3187.675960] other info that might help us debug this:
[ 3187.690459]  Possible unsafe locking scenario:
[ 3187.690459] 
[ 3187.704224]        CPU0
[ 3187.710485]        ----
[ 3187.716711]   lock(&power_domains->lock);
[ 3187.724718]   lock(&power_domains->lock);
[ 3187.732663]
[ 3187.732663]  *** DEADLOCK ***
[ 3187.732663]
[ 3187.749460]  May be due to missing lock nesting notation [ 3187.749460] [ 3187.763833] 5 locks held by Xorg/32212:
[ 3187.771523]  #0:  (drm_global_mutex){+.+.+.}, at: [<ffffffffa02ad5c8>] drm_release+0x3b/0x49b [drm] [ 3187.785216]  #1:  (&dev->mode_config.mutex){+.+.+.}, at: [<ffffffffa02c4d85>] drm_modeset_lock_all+0x54/0xcd [drm] [ 3187.800437]  #2:  (crtc_ww_class_acquire){+.+.+.}, at: [<ffffffffa02c4d8f>] drm_modeset_lock_all+0x5e/0xcd [drm] [ 3187.815488]  #3:  (crtc_ww_class_mutex){+.+.+.}, at: [<ffffffffa02c46b0>] drm_modeset_lock+0x75/0xfc [drm] [ 3187.830094]  #4:  (&power_domains->lock){+.+...}, at: [<ffffffffa036aced>] intel_display_power_get+0x38/0xcb [i915] [ 3187.845534] [ 3187.845534] stack backtrace:
[ 3187.857685] CPU: 2 PID: 32212 Comm: Xorg Tainted: G     U  W       4.3.0-rc4-bsw+ #2488
[ 3187.870331] Hardware name: Intel Corporation CHERRYVIEW C0 PLATFORM/Braswell CRB, BIOS BRAS.X64.B085.R00.1509110553 09/11/2015 [ 3187.886827]  0000000000000000 ffff880175eff8e0 ffffffff8128d59e ffffffff823f5ee0 [ 3187.898904]  ffff880175eff958 ffffffff810a7a08 0000000000000000 ffff880179d1c5d0 [ 3187.910954]  0000000000000004 0000000000000006 45422a91588a4c3e 0000000000000005 [ 3187.923011] Call Trace:
[ 3187.929451]  [<ffffffff8128d59e>] dump_stack+0x4e/0x79 [ 3187.938912]  [<ffffffff810a7a08>] __lock_acquire+0x7ab/0x12af [ 3187.949027]  [<ffffffff810a8d04>] lock_acquire+0x10e/0x1a9 [ 3187.958859]  [<ffffffffa036aced>] ? intel_display_power_get+0x38/0xcb [i915] [ 3187.970476]  [<ffffffffa036aced>] ? intel_display_power_get+0x38/0xcb [i915] [ 3187.982011]  [<ffffffff814df4a3>] mutex_lock_nested+0x71/0x346 [ 3187.992167]  [<ffffffffa036aced>] ? intel_display_power_get+0x38/0xcb [i915] [ 3188.003655]  [<ffffffff814e20ee>] ? _raw_spin_unlock_irqrestore+0x4b/0x60
[ 3188.014829]  [<ffffffff81364565>] ? __pm_runtime_resume+0x71/0x7e [ 3188.025269]  [<ffffffffa036aced>] intel_display_power_get+0x38/0xcb [i915] [ 3188.036544]  [<ffffffffa036aced>] ? intel_display_power_get+0x38/0xcb [i915] [ 3188.047968]  [<ffffffffa03e8cc4>] intel_hdmi_set_edid+0x3f/0xd6 [i915] [ 3188.058766]  [<ffffffffa03e8e1a>] intel_hdmi_hot_plug+0xbf/0xfb [i915] [ 3188.069484]  [<ffffffffa03c9e78>] intel_hpd_init+0xfa/0x10b [i915] [ 3188.079753]  [<ffffffffa036968d>] vlv_display_power_well_init+0xdb/0xe8 [i915] [ 3188.091224]  [<ffffffffa0369763>] chv_pipe_power_well_enable+0x62/0x67 [i915] [ 3188.102594]  [<ffffffffa036ad55>] intel_display_power_get+0xa0/0xcb [i915] [ 3188.113657]  [<ffffffffa03bdd7a>] modeset_get_crtc_power_domains+0x11d/0x13c [i915] [ 3188.125589]  [<ffffffffa03bdfc1>] intel_atomic_commit+0x228/0xf1b [i915] [ 3188.136522]  [<ffffffffa02c586c>] ? drm_atomic_check_only+0x37b/0x4da [drm] [ 3188.147676]  [<ffffffffa02c5a18>] drm_atomic_commit+0x4d/0x52 [drm] [ 3188.158023]  [<ffffffffa032fa25>] restore_fbdev_mode+0x11e/0x286 [drm_kms_helper] [ 3188.169734]  [<ffffffffa0331443>] drm_fb_helper_restore_fbdev_mode_unlocked+0x36/0x6d [drm_kms_helper] [ 3188.183557]  [<ffffffffa03d270f>] intel_fbdev_restore_mode+0x22/0x7a [i915] [ 3188.194701]  [<ffffffffa03f70b8>] i915_driver_lastclose+0xe/0x15 [i915] [ 3188.205361]  [<ffffffffa02ad4c2>] drm_lastclose+0x3f/0x10a [drm] [ 3188.215323]  [<ffffffffa02ad9ee>] drm_release+0x461/0x49b [drm] [ 3188.225195]  [<ffffffff8118577c>] __fput+0x100/0x1b3 [ 3188.233956]  [<ffffffff81185865>] ____fput+0xe/0x10 [ 3188.242594]  [<ffffffff81084540>] task_work_run+0x6a/0x93 [ 3188.251816]  [<ffffffff810015d6>] prepare_exit_to_usermode+0x9e/0xaf
[ 3188.262111]  [<ffffffff810017d6>] syscall_return_slowpath+0x1ef/0x264
[ 3188.272486]  [<ffffffff81084462>] ? task_work_add+0x44/0x53 [ 3188.281885]  [<ffffffff811858e3>] ? fput+0x7c/0x83 [ 3188.290389]  [<ffffffff810a6e52>] ? trace_hardirqs_on_caller+0x16/0x196
[ 3188.300941]  [<ffffffff81000c87>] ? trace_hardirqs_on_thunk+0x17/0x19 [ 3188.311245]  [<ffffffff814e2ab1>] int_ret_from_sys_call+0x25/0x9f
Quoted 19 lines
drivers/gpu/drm/i915/intel_hotplug.c +0 -1
 }
 
 void intel_hpd_init_work(struct drm_i915_private *dev_priv) diff 
--git a/drivers/gpu/drm/i915/intel_sdvo.c

b/drivers/gpu/drm/i915/intel_sdvo.c
index 05521b5..55859e9 100644

drivers/gpu/drm/i915/intel_sdvo.c +0 -2
@@ -2433,7 +2433,6 @@ intel_sdvo_dvi_init(struct intel_sdvo *intel_sdvo, int device)
 		 * Ensure that they get re-enabled when an interrupt happens.
 		 */
 		intel_encoder->hot_plug = intel_sdvo_enable_hotplug;
-		intel_sdvo_enable_hotplug(intel_encoder);
 	} else {
 		intel_connector->polled = DRM_CONNECTOR_POLL_CONNECT | DRM_CONNECTOR_POLL_DISCONNECT;
 	}
--

1.7.10.4

Quoted 4 lines

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

--
Ville Syrjälä
Intel OTC
_______________________________________________
Intel-gfx mailing list
[email protected]
http://lists.freedesktop.org/mailman/listinfo/intel-gfx
_______________________________________________
Intel-gfx mailing list
[email protected]
http://lists.freedesktop.org/mailman/listinfo/intel-gfx

#32 lore raw

PATCH 3/6 drm/i915: Make intel_digital_port_connected global

In reply to Sonika Jindal

This is to allow live status check for HDMI as well. Also, using intel_encoder->hpd_pin to check the live status for bxt because of BXT A0/A1 WA for HPD pins.

---
 drivers/gpu/drm/i915/intel_dp.c  |   11 +++++++----
 drivers/gpu/drm/i915/intel_drv.h |    2 ++
 2 files changed, 9 insertions(+), 4 deletions(-)
drivers/gpu/drm/i915/intel_dp.c +7 -4
index 796f930..fedf6d1 100644
@@ -4663,11 +4663,14 @@ static bool vlv_digital_port_connected(struct drm_i915_private *dev_priv,
 }
 
 static bool bxt_digital_port_connected(struct drm_i915_private *dev_priv,
-				       struct intel_digital_port *port)
+				       struct intel_digital_port *intel_dig_port)
 {
+	struct intel_encoder *intel_encoder = &intel_dig_port->base;
+	enum port port;
 	u32 bit;
 
-	switch (port->port) {
+	intel_hpd_pin_to_port(intel_encoder->hpd_pin, &port);
+	switch (port) {
 	case PORT_A:
 		bit = BXT_DE_PORT_HP_DDIA;
 		break;
@@ -4678,7 +4681,7 @@ static bool bxt_digital_port_connected(struct drm_i915_private *dev_priv,
 		bit = BXT_DE_PORT_HP_DDIC;
 		break;
 	default:
-		MISSING_CASE(port->port);
+		MISSING_CASE(port);
 		return false;
 	}
 
@@ -4692,7 +4695,7 @@ static bool bxt_digital_port_connected(struct drm_i915_private *dev_priv,
  *
  * Return %true if @port is connected, %false otherwise.
  */
-static bool intel_digital_port_connected(struct drm_i915_private *dev_priv,
+bool intel_digital_port_connected(struct drm_i915_private *dev_priv,
 					 struct intel_digital_port *port)
 {
 	if (HAS_PCH_IBX(dev_priv))
drivers/gpu/drm/i915/intel_drv.h +2 -0
index b6c2c20..ac6d748 100644
@@ -1210,6 +1210,8 @@ void intel_edp_drrs_disable(struct intel_dp *intel_dp);
 void intel_edp_drrs_invalidate(struct drm_device *dev,
 		unsigned frontbuffer_bits);
 void intel_edp_drrs_flush(struct drm_device *dev, unsigned frontbuffer_bits);
+bool intel_digital_port_connected(struct drm_i915_private *dev_priv,
+					 struct intel_digital_port *port);
 
 /* intel_dp_mst.c */
 int intel_dp_mst_encoder_init(struct intel_digital_port *intel_dig_port, int conn_id);
Signature · 6 lines
1.7.10.4

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

Re: [PATCH 3/6] drm/i915: Make intel_digital_port_connected global

Quoted 4 lines · Sonika Jindal

This is to allow live status check for HDMI as well. Also, using intel_encoder->hpd_pin to check the live status for bxt because of BXT A0/A1 WA for HPD pins.

This is actually the only thing this patch does... So please use separated patch for this....

But the global intel_digital_port connected can be done in the same patch that uses it actually...

Quoted 72 lines
---
 drivers/gpu/drm/i915/intel_dp.c  |   11 +++++++----
 drivers/gpu/drm/i915/intel_drv.h |    2 ++
 2 files changed, 9 insertions(+), 4 deletions(-)
drivers/gpu/drm/i915/intel_dp.c +5 -2
index 796f930..fedf6d1 100644
@@ -4663,11 +4663,14 @@ static bool vlv_digital_port_connected(struct
 drm_i915_private *dev_priv,
 }
 
 static bool bxt_digital_port_connected(struct drm_i915_private *dev_priv,
-                                      struct intel_digital_port *port)
+                                      struct intel_digital_port
 *intel_dig_port)
 {
+       struct intel_encoder *intel_encoder = &intel_dig_port->base;
+       enum port port;
        u32 bit;
 
-       switch (port->port) {
+       intel_hpd_pin_to_port(intel_encoder->hpd_pin, &port);
+       switch (port) {
        case PORT_A:
                bit = BXT_DE_PORT_HP_DDIA;
                break;
drivers/gpu/drm/i915/intel_dp.c +2 -2
@@ -4678,7 +4681,7 @@ static bool bxt_digital_port_connected(struct
 drm_i915_private *dev_priv,
                bit = BXT_DE_PORT_HP_DDIC;
                break;
        default:
-               MISSING_CASE(port->port);
+               MISSING_CASE(port);
                return false;
        }
@@ -4692,7 +4695,7 @@ static bool bxt_digital_port_connected(struct
 drm_i915_private *dev_priv,
  *
  * Return %true if @port is connected, %false otherwise.
  */
-static bool intel_digital_port_connected(struct drm_i915_private
 *dev_priv,
+bool intel_digital_port_connected(struct drm_i915_private *dev_priv,
                                         struct intel_digital_port *port)
 {
        if (HAS_PCH_IBX(dev_priv))
drivers/gpu/drm/i915/intel_drv.h +2 -0
index b6c2c20..ac6d748 100644
@@ -1210,6 +1210,8 @@ void intel_edp_drrs_disable(struct intel_dp
 *intel_dp);
 void intel_edp_drrs_invalidate(struct drm_device *dev,
                unsigned frontbuffer_bits);
 void intel_edp_drrs_flush(struct drm_device *dev, unsigned
 frontbuffer_bits);
+bool intel_digital_port_connected(struct drm_i915_private *dev_priv,
+                                        struct intel_digital_port *port);
 

/* intel_dp_mst.c */
int intel_dp_mst_encoder_init(struct intel_digital_port *intel_dig_port,
int conn_id);
--
1.7.10.4

Quoted 5 lines

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

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

#34 lore raw

PATCH 1/1 drm/i915/bxt: Use intel_encoder->hpd_pin to check live status

Using intel_encoder's hpd_pin to check the live status because of BXT A0/A1 WA for HPD pins and hpd_pin contains the updated pin for the corresponding port.

---
 drivers/gpu/drm/i915/intel_dp.c |    9 ++++++---
 1 file changed, 6 insertions(+), 3 deletions(-)
drivers/gpu/drm/i915/intel_dp.c +6 -3
index 796f930..bf17030 100644
@@ -4663,11 +4663,14 @@ static bool vlv_digital_port_connected(struct drm_i915_private *dev_priv,
 }
 
 static bool bxt_digital_port_connected(struct drm_i915_private *dev_priv,
-				       struct intel_digital_port *port)
+				       struct intel_digital_port *intel_dig_port)
 {
+	struct intel_encoder *intel_encoder = &intel_dig_port->base;
+	enum port port;
 	u32 bit;
 
-	switch (port->port) {
+	intel_hpd_pin_to_port(intel_encoder->hpd_pin, &port);
+	switch (port) {
 	case PORT_A:
 		bit = BXT_DE_PORT_HP_DDIA;
 		break;
@@ -4678,7 +4681,7 @@ static bool bxt_digital_port_connected(struct drm_i915_private *dev_priv,
 		bit = BXT_DE_PORT_HP_DDIC;
 		break;
 	default:
-		MISSING_CASE(port->port);
+		MISSING_CASE(port);
 		return false;
 	}
 
Signature · 6 lines
1.7.10.4

_______________________________________________
Intel-gfx mailing list
[email protected]
http://lists.freedesktop.org/mailman/listinfo/intel-gfx
Reviewed-by on v1 1/1
#35 lore raw

Re: [PATCH] drm/i915/bxt: Use intel_encoder->hpd_pin to check live status

Quoted 4 lines · Sonika Jindal

Using intel_encoder's hpd_pin to check the live status because of BXT A0/A1 WA for HPD pins and hpd_pin contains the updated pin for the corresponding port.

It makes sense, but is it always true? or we should have a fallback to intel_dig_port->port in some case or in some stepping after A0/A1?

With this clarified feel free to use:

Quoted 43 lines
---
 drivers/gpu/drm/i915/intel_dp.c |    9 ++++++---
 1 file changed, 6 insertions(+), 3 deletions(-)
drivers/gpu/drm/i915/intel_dp.c +5 -2
index 796f930..bf17030 100644
@@ -4663,11 +4663,14 @@ static bool vlv_digital_port_connected(struct
 drm_i915_private *dev_priv,
 }
 
 static bool bxt_digital_port_connected(struct drm_i915_private *dev_priv,
-                                      struct intel_digital_port *port)
+                                      struct intel_digital_port
 *intel_dig_port)
 {
+       struct intel_encoder *intel_encoder = &intel_dig_port->base;
+       enum port port;
        u32 bit;
 
-       switch (port->port) {
+       intel_hpd_pin_to_port(intel_encoder->hpd_pin, &port);
+       switch (port) {
        case PORT_A:
                bit = BXT_DE_PORT_HP_DDIA;
                break;
drivers/gpu/drm/i915/intel_dp.c +1 -2
@@ -4678,7 +4681,7 @@ static bool bxt_digital_port_connected(struct
 drm_i915_private *dev_priv,
                bit = BXT_DE_PORT_HP_DDIC;
                break;
        default:
-               MISSING_CASE(port->port);
+               MISSING_CASE(port);
                return false;
        }
 
--

1.7.10.4

Quoted 5 lines

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

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

author Reviewed-by Rodrigo Vivi
#36 lore raw

Re: [PATCH] drm/i915/bxt: Use intel_encoder->hpd_pin to check live status

Thanks Rodrigo.
This is always true. The hpd_pin is the right pin which we set for particular port. So no fallback needed.

Regards,
Sonika

From: Rodrigo Vivi [mailto:[email protected]]
Sent: Friday, September 11, 2015 11:31 PM
To: Jindal, Sonika; [email protected]
Subject: Re: [Intel-gfx] [PATCH] drm/i915/bxt: Use intel_encoder->hpd_pin to check live status

On Fri, Sep 11, 2015 at 4:40 AM Sonika Jindal <[email protected]<mailto:[email protected]>> wrote:
Using intel_encoder's hpd_pin to check the live status
because of BXT A0/A1 WA for HPD pins and hpd_pin contains the
updated pin for the corresponding port.

It makes sense, but is it always true? or we should have a fallback to intel_dig_port->port in some case or in some stepping after A0/A1?

With this clarified feel free to use:

---
 drivers/gpu/drm/i915/intel_dp.c |    9 ++++++---
 1 file changed, 6 insertions(+), 3 deletions(-)
drivers/gpu/drm/i915/intel_dp.c +6 -3
index 796f930..bf17030 100644
@@ -4663,11 +4663,14 @@ static bool vlv_digital_port_connected(struct drm_i915_private *dev_priv,
 }
 
 static bool bxt_digital_port_connected(struct drm_i915_private *dev_priv,
-                                      struct intel_digital_port *port)
+                                      struct intel_digital_port *intel_dig_port)
 {
+       struct intel_encoder *intel_encoder = &intel_dig_port->base;
+       enum port port;
        u32 bit;
 
-       switch (port->port) {
+       intel_hpd_pin_to_port(intel_encoder->hpd_pin, &port);
+       switch (port) {
        case PORT_A:
                bit = BXT_DE_PORT_HP_DDIA;
                break;
@@ -4678,7 +4681,7 @@ static bool bxt_digital_port_connected(struct drm_i915_private *dev_priv,
                bit = BXT_DE_PORT_HP_DDIC;
                break;
        default:
-               MISSING_CASE(port->port);
+               MISSING_CASE(port);
                return false;
        }
 

--
1.7.10.4

_______________________________________________
Intel-gfx mailing list [email protected]<mailto:[email protected]>
http://lists.freedesktop.org/mailman/listinfo/intel-gfx

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

#37 lore raw

Re: [PATCH] drm/i915/bxt: Use intel_encoder->hpd_pin to check live status

In reply to Sonika Jindal
Quoted 16 lines · Sonika Jindal

Using intel_encoder's hpd_pin to check the live status because of BXT A0/A1 WA for HPD pins and hpd_pin contains the updated pin for the corresponding port.

---
 drivers/gpu/drm/i915/intel_dp.c |    9 ++++++---
 1 file changed, 6 insertions(+), 3 deletions(-)
drivers/gpu/drm/i915/intel_dp.c +0 -0
index 796f930..bf17030 100644
@@ -4663,11 +4663,14 @@ static bool vlv_digital_port_connected(struct drm_i915_private *dev_priv,
 }
 

drivers/gpu/drm/i915/intel_dp.c static bool vlv_digital_port_connected(struct drm_i915_private *dev_priv, · line 4672

Quoted 10 lines
drivers/gpu/drm/i915/intel_dp.c +4 -2
 static bool bxt_digital_port_connected(struct drm_i915_private *dev_priv,
-				       struct intel_digital_port *port)
+				       struct intel_digital_port *intel_dig_port)
 {
+	struct intel_encoder *intel_encoder = &intel_dig_port->base;
+	enum port port;
 	u32 bit;
 
-	switch (port->port) {
+	intel_hpd_pin_to_port(intel_encoder->hpd_pin, &port);

This looks very wrong - the function you're calling here maps from hpd_pin to the port, but what you actually want is to map from port to pin. That mapping is in intel_encoder->hpd_pin.
-Daniel

Quoted 20 lines
drivers/gpu/drm/i915/intel_dp.c +2 -1
+	switch (port) {
 	case PORT_A:
 		bit = BXT_DE_PORT_HP_DDIA;
 		break;
@@ -4678,7 +4681,7 @@ static bool bxt_digital_port_connected(struct drm_i915_private *dev_priv,
 		bit = BXT_DE_PORT_HP_DDIC;
 		break;
 	default:
-		MISSING_CASE(port->port);
+		MISSING_CASE(port);
 		return false;
 	}
 
Signature · 6 lines
1.7.10.4

_______________________________________________
Intel-gfx mailing list
[email protected]
http://lists.freedesktop.org/mailman/listinfo/intel-gfx
Signature · 7 lines
Daniel Vetter
Software Engineer, Intel Corporation
http://blog.ffwll.ch
_______________________________________________
Intel-gfx mailing list
[email protected]
http://lists.freedesktop.org/mailman/listinfo/intel-gfx
#38 lore raw

Re: [PATCH] drm/i915/bxt: Use intel_encoder->hpd_pin to check live status

Quoted 25 lines · Daniel Vetter
Quoted 24 lines · Sonika Jindal

Using intel_encoder's hpd_pin to check the live status because of BXT A0/A1 WA for HPD pins and hpd_pin contains the updated pin for the corresponding port.

---
  drivers/gpu/drm/i915/intel_dp.c |    9 ++++++---
  1 file changed, 6 insertions(+), 3 deletions(-)
drivers/gpu/drm/i915/intel_dp.c +3 -1
index 796f930..bf17030 100644
@@ -4663,11 +4663,14 @@ static bool vlv_digital_port_connected(struct drm_i915_private *dev_priv,
  }
 
  static bool bxt_digital_port_connected(struct drm_i915_private *dev_priv,
-				       struct intel_digital_port *port)
+				       struct intel_digital_port *intel_dig_port)
  {
+	struct intel_encoder *intel_encoder = &intel_dig_port->base;
+	enum port port;
  	u32 bit;
 
Quoted 7 lines

drivers/gpu/drm/i915/intel_dp.c static bool vlv_digital_port_connected(struct drm_i915_private *dev_priv, · line 4672

Quoted 2 lines
drivers/gpu/drm/i915/intel_dp.c +1 -1
-	switch (port->port) {
+	intel_hpd_pin_to_port(intel_encoder->hpd_pin, &port);

This looks very wrong - the function you're calling here maps from hpd_pin to the port, but what you actually want is to map from port to pin. That mapping is in intel_encoder->hpd_pin.
-Daniel

No, I want to find out the correct port whose hpd bit we need to check.
That port we are finding based upon the hpd_pin we have set for that encoder. This was the whole point of using hpd_pin for the BXT A0/A1 WA. That instead of checking different HPD bit based upon the port, just change the hpd_pin itself in intel_encoder.

Regards,
Sonika

Quoted 14 lines
Quoted 14 lines
drivers/gpu/drm/i915/intel_dp.c +2 -1
+	switch (port) {
  	case PORT_A:
  		bit = BXT_DE_PORT_HP_DDIA;
  		break;
@@ -4678,7 +4681,7 @@ static bool bxt_digital_port_connected(struct drm_i915_private *dev_priv,
  		bit = BXT_DE_PORT_HP_DDIC;
  		break;
  	default:
-		MISSING_CASE(port->port);
+		MISSING_CASE(port);
  		return false;
  	}
 
Quoted 8 lines
Quoted 7 lines

--
1.7.10.4

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

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

#39 lore raw

Re: [PATCH] drm/i915/bxt: Use intel_encoder->hpd_pin to check live status

Quoted 35 lines · Sonika
Quoted 25 lines · Daniel Vetter
Quoted 24 lines · Sonika Jindal

Using intel_encoder's hpd_pin to check the live status because of BXT A0/A1 WA for HPD pins and hpd_pin contains the updated pin for the corresponding port.

---
 drivers/gpu/drm/i915/intel_dp.c |    9 ++++++---
 1 file changed, 6 insertions(+), 3 deletions(-)
drivers/gpu/drm/i915/intel_dp.c +3 -1
index 796f930..bf17030 100644
@@ -4663,11 +4663,14 @@ static bool vlv_digital_port_connected(struct drm_i915_private *dev_priv,
 }
 
 static bool bxt_digital_port_connected(struct drm_i915_private *dev_priv,
-				       struct intel_digital_port *port)
+				       struct intel_digital_port *intel_dig_port)
 {
+	struct intel_encoder *intel_encoder = &intel_dig_port->base;
+	enum port port;
 	u32 bit;
 
Quoted 7 lines

drivers/gpu/drm/i915/intel_dp.c static bool vlv_digital_port_connected(struct drm_i915_private *dev_priv, · line 4672

Quoted 2 lines
drivers/gpu/drm/i915/intel_dp.c +1 -1
-	switch (port->port) {
+	intel_hpd_pin_to_port(intel_encoder->hpd_pin, &port);

This looks very wrong - the function you're calling here maps from hpd_pin to the port, but what you actually want is to map from port to pin. That mapping is in intel_encoder->hpd_pin.
-Daniel

Quoted 5 lines

No, I want to find out the correct port whose hpd bit we need to check. That port we are finding based upon the hpd_pin we have set for that encoder. This was the whole point of using hpd_pin for the BXT A0/A1 WA. That instead of checking different HPD bit based upon the port, just change the hpd_pin itself in intel_encoder.

Ah right I was getting confused. Patch applied now, thanks for clarifying. -Daniel

Signature · 7 lines
Daniel Vetter
Software Engineer, Intel Corporation
http://blog.ffwll.ch
_______________________________________________
Intel-gfx mailing list
[email protected]
http://lists.freedesktop.org/mailman/listinfo/intel-gfx
#40 lore raw

PATCH 4/6 drm/i915: drm/i915: Check live status before reading edid

In reply to Sonika Jindal

The Bspec is very clear that Live status must be checked about before trying to read EDID over DDC channel. This patch makes sure that HDMI EDID is read only when live status us up.

The live status doesn't seem to perform very consistent across various platforms when tested with different monitors. The reason behind that is some monitors are late to provide right voltage to set live_status up. So, after getting the interrupt, for a small duration, live status reg fluctuates, and then settles down showing the correct staus.

This is explained here in, in a rough way:
HPD line ________________

			 |\ T1 = Monitor Hotplug causing IRQ
			 | \______________________________________
			 | |
                         | |
			 | |   T2 = Live status is stable 
			 | |  _____________________________________
			 | | /|
Live status _____________|_|/ |
			 | |  |
			 | |  |
			 | |  |
			T0 T1  T2

(Between T1 and T2 Live status fluctuates or can be even low, depending on
the monitor)

After several experiments, we have concluded that a max delay of 30ms is enough to allow the live status to settle down with most of the monitors. This total delay of 30ms has been split into a resolution of 3 retries of 10ms each, for the better cases.

This delay is kept at 30ms, keeping in consideration that, HDCP compliance expect the HPD handler to respond a plug out in 100ms, by disabling the port.

v2: Adding checks for VLV/CHV as well. Reusing old ibx and g4x functions to check digital port status. Adding a separate function to get bxt live status (Daniel)
v3: Using intel_encoder->hpd_pin to check the live status (Siva) Moving the live status read to intel_hdmi_probe and passing parameter to read/not to read the edid. (me)
v4:
* Added live status check for all platforms using intel_digital_port_connected.
* Rebased on top of Jani's DP cleanup series
* Some monitors take time in setting the live status. So retry for few times if this is a connect HPD

---
 drivers/gpu/drm/i915/intel_hdmi.c |   35 ++++++++++++++++++++++++++++-------
 1 file changed, 28 insertions(+), 7 deletions(-)
drivers/gpu/drm/i915/intel_hdmi.c +28 -7
index 1eda71a..d82887b 100644
@@ -1329,22 +1329,23 @@ intel_hdmi_unset_edid(struct drm_connector *connector)
 }
 
 static bool
-intel_hdmi_set_edid(struct drm_connector *connector)
+intel_hdmi_set_edid(struct drm_connector *connector, bool force)
 {
 	struct drm_i915_private *dev_priv = to_i915(connector->dev);
 	struct intel_hdmi *intel_hdmi = intel_attached_hdmi(connector);
 	struct intel_encoder *intel_encoder =
 		&hdmi_to_dig_port(intel_hdmi)->base;
 	enum intel_display_power_domain power_domain;
-	struct edid *edid;
+	struct edid *edid = NULL;
 	bool connected = false;
 
 	power_domain = intel_display_port_power_domain(intel_encoder);
 	intel_display_power_get(dev_priv, power_domain);
 
-	edid = drm_get_edid(connector,
-			    intel_gmbus_get_adapter(dev_priv,
-						    intel_hdmi->ddc_bus));
+	if (force)
+		edid = drm_get_edid(connector,
+				    intel_gmbus_get_adapter(dev_priv,
+				    intel_hdmi->ddc_bus));
 
 	intel_display_power_put(dev_priv, power_domain);
 
@@ -1374,6 +1375,26 @@ void intel_hdmi_probe(struct intel_encoder *intel_encoder)
 			enc_to_intel_hdmi(&intel_encoder->base);
 	struct intel_connector *intel_connector =
 				intel_hdmi->attached_connector;
+	struct drm_i915_private *dev_priv = to_i915(intel_encoder->base.dev);
+	bool live_status = false;
+	unsigned int retry = 3;
+
+	live_status = intel_digital_port_connected(dev_priv,
+						   hdmi_to_dig_port(intel_hdmi));
+	if (!intel_connector->detect_edid && live_status == false) {
+		/*
+		 * Hotplug had occurred and old status was disconnected,
+		 * so it might be possible that live status is not set,
+		 * so retry for few times
+		 */
+		do {
+			mdelay(10);
+			live_status = intel_digital_port_connected(dev_priv,
+						hdmi_to_dig_port(intel_hdmi));
+			if (live_status)
+				break;
+		} while (retry--);
+	}
 
 	/*
 	 * We are here, means there is a hotplug or a force
@@ -1381,7 +1402,7 @@ void intel_hdmi_probe(struct intel_encoder *intel_encoder)
 	 * DDC bus to check the current status of HDMI.
 	 */
 	intel_hdmi_unset_edid(&intel_connector->base);
-	if (intel_hdmi_set_edid(&intel_connector->base))
+	if (intel_hdmi_set_edid(&intel_connector->base, live_status))
 		DRM_DEBUG_DRIVER("DDC probe: got EDID\n");
 	else
 		DRM_DEBUG_DRIVER("DDC probe: no EDID\n");
@@ -1432,7 +1453,7 @@ intel_hdmi_force(struct drm_connector *connector)
 	if (connector->status != connector_status_connected)
 		return;
 
-	intel_hdmi_set_edid(connector);
+	intel_hdmi_set_edid(connector, true);
 	hdmi_to_dig_port(intel_hdmi)->base.type = INTEL_OUTPUT_HDMI;
 }
 
Signature · 6 lines
1.7.10.4

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

Re: [PATCH 4/6] drm/i915: drm/i915: Check live status before reading edid

Quoted 49 lines · Sonika Jindal

The Bspec is very clear that Live status must be checked about before trying to read EDID over DDC channel. This patch makes sure that HDMI EDID is read only when live status us up.

The live status doesn't seem to perform very consistent across various platforms when tested with different monitors. The reason behind that is some monitors are late to provide right voltage to set live_status up. So, after getting the interrupt, for a small duration, live status reg fluctuates, and then settles down showing the correct staus.

This is explained here in, in a rough way:
HPD line ________________

			 |\ T1 = Monitor Hotplug causing IRQ
			 | \______________________________________
			 | |
                         | |
			 | |   T2 = Live status is stable 
			 | |  _____________________________________
			 | | /|
Live status _____________|_|/ |
			 | |  |
			 | |  |
			 | |  |
			T0 T1  T2

(Between T1 and T2 Live status fluctuates or can be even low, depending on
the monitor)

After several experiments, we have concluded that a max delay of 30ms is enough to allow the live status to settle down with most of the monitors. This total delay of 30ms has been split into a resolution of 3 retries of 10ms each, for the better cases.

This delay is kept at 30ms, keeping in consideration that, HDCP compliance expect the HPD handler to respond a plug out in 100ms, by disabling the port.

v2: Adding checks for VLV/CHV as well. Reusing old ibx and g4x functions to check digital port status. Adding a separate function to get bxt live status (Daniel)
v3: Using intel_encoder->hpd_pin to check the live status (Siva) Moving the live status read to intel_hdmi_probe and passing parameter to read/not to read the edid. (me)
v4:
* Added live status check for all platforms using intel_digital_port_connected.
* Rebased on top of Jani's DP cleanup series
* Some monitors take time in setting the live status. So retry for few times if this is a connect HPD

Quoted 1 line

Really pretty commit message! Had some minor comments on two other patches in this series while reading through them, but looks good otherwise. Please sign up someone for final review of the details and I'll pull in the patches once that's done.
-Daniel

Quoted 89 lines
---
 drivers/gpu/drm/i915/intel_hdmi.c |   35 ++++++++++++++++++++++++++++-------
 1 file changed, 28 insertions(+), 7 deletions(-)
drivers/gpu/drm/i915/intel_hdmi.c +28 -7
index 1eda71a..d82887b 100644
@@ -1329,22 +1329,23 @@ intel_hdmi_unset_edid(struct drm_connector *connector)
 }
 
 static bool
-intel_hdmi_set_edid(struct drm_connector *connector)
+intel_hdmi_set_edid(struct drm_connector *connector, bool force)
 {
 	struct drm_i915_private *dev_priv = to_i915(connector->dev);
 	struct intel_hdmi *intel_hdmi = intel_attached_hdmi(connector);
 	struct intel_encoder *intel_encoder =
 		&hdmi_to_dig_port(intel_hdmi)->base;
 	enum intel_display_power_domain power_domain;
-	struct edid *edid;
+	struct edid *edid = NULL;
 	bool connected = false;
 
 	power_domain = intel_display_port_power_domain(intel_encoder);
 	intel_display_power_get(dev_priv, power_domain);
 
-	edid = drm_get_edid(connector,
-			    intel_gmbus_get_adapter(dev_priv,
-						    intel_hdmi->ddc_bus));
+	if (force)
+		edid = drm_get_edid(connector,
+				    intel_gmbus_get_adapter(dev_priv,
+				    intel_hdmi->ddc_bus));
 
 	intel_display_power_put(dev_priv, power_domain);
 
@@ -1374,6 +1375,26 @@ void intel_hdmi_probe(struct intel_encoder *intel_encoder)
 			enc_to_intel_hdmi(&intel_encoder->base);
 	struct intel_connector *intel_connector =
 				intel_hdmi->attached_connector;
+	struct drm_i915_private *dev_priv = to_i915(intel_encoder->base.dev);
+	bool live_status = false;
+	unsigned int retry = 3;
+
+	live_status = intel_digital_port_connected(dev_priv,
+						   hdmi_to_dig_port(intel_hdmi));
+	if (!intel_connector->detect_edid && live_status == false) {
+		/*
+		 * Hotplug had occurred and old status was disconnected,
+		 * so it might be possible that live status is not set,
+		 * so retry for few times
+		 */
+		do {
+			mdelay(10);
+			live_status = intel_digital_port_connected(dev_priv,
+						hdmi_to_dig_port(intel_hdmi));
+			if (live_status)
+				break;
+		} while (retry--);
+	}
 
 	/*
 	 * We are here, means there is a hotplug or a force
@@ -1381,7 +1402,7 @@ void intel_hdmi_probe(struct intel_encoder *intel_encoder)
 	 * DDC bus to check the current status of HDMI.
 	 */
 	intel_hdmi_unset_edid(&intel_connector->base);
-	if (intel_hdmi_set_edid(&intel_connector->base))
+	if (intel_hdmi_set_edid(&intel_connector->base, live_status))
 		DRM_DEBUG_DRIVER("DDC probe: got EDID\n");
 	else
 		DRM_DEBUG_DRIVER("DDC probe: no EDID\n");
@@ -1432,7 +1453,7 @@ intel_hdmi_force(struct drm_connector *connector)
 	if (connector->status != connector_status_connected)
 		return;
 
-	intel_hdmi_set_edid(connector);
+	intel_hdmi_set_edid(connector, true);
 	hdmi_to_dig_port(intel_hdmi)->base.type = INTEL_OUTPUT_HDMI;
 }
 
Signature · 6 lines
1.7.10.4

_______________________________________________
Intel-gfx mailing list
[email protected]
http://lists.freedesktop.org/mailman/listinfo/intel-gfx
Signature · 7 lines
Daniel Vetter
Software Engineer, Intel Corporation
http://blog.ffwll.ch
_______________________________________________
Intel-gfx mailing list
[email protected]
http://lists.freedesktop.org/mailman/listinfo/intel-gfx
#42 lore raw

PATCH 1/1 drm/i915: Check live status before reading edid

Adding this for SKL onwards.

v2: Adding checks for VLV/CHV as well. Reusing old ibx and g4x functions to check digital port status. Adding a separate function to get bxt live status (Daniel)
v3: Using intel_encoder->hpd_pin to check the live status (Siva) Moving the live status read to intel_hdmi_probe and passing parameter to read/not to read the edid. (me)
v4:
* Added live status check for all platforms using intel_digital_port_connected.
* Rebased on top of Jani's DP cleanup series
* Some monitors take time in setting the live status. So retry for few times if this is a connect HPD
v5: Removed extra "drm/i915" from commit message. Adding Shashank's sob which was missed.

---
 drivers/gpu/drm/i915/intel_hdmi.c |   35 ++++++++++++++++++++++++++++-------
 1 file changed, 28 insertions(+), 7 deletions(-)
drivers/gpu/drm/i915/intel_hdmi.c +28 -7
index 1eda71a..d82887b 100644
@@ -1329,22 +1329,23 @@ intel_hdmi_unset_edid(struct drm_connector *connector)
 }
 
 static bool
-intel_hdmi_set_edid(struct drm_connector *connector)
+intel_hdmi_set_edid(struct drm_connector *connector, bool force)
 {
 	struct drm_i915_private *dev_priv = to_i915(connector->dev);
 	struct intel_hdmi *intel_hdmi = intel_attached_hdmi(connector);
 	struct intel_encoder *intel_encoder =
 		&hdmi_to_dig_port(intel_hdmi)->base;
 	enum intel_display_power_domain power_domain;
-	struct edid *edid;
+	struct edid *edid = NULL;
 	bool connected = false;
 
 	power_domain = intel_display_port_power_domain(intel_encoder);
 	intel_display_power_get(dev_priv, power_domain);
 
-	edid = drm_get_edid(connector,
-			    intel_gmbus_get_adapter(dev_priv,
-						    intel_hdmi->ddc_bus));
+	if (force)
+		edid = drm_get_edid(connector,
+				    intel_gmbus_get_adapter(dev_priv,
+				    intel_hdmi->ddc_bus));
 
 	intel_display_power_put(dev_priv, power_domain);
 
@@ -1374,6 +1375,26 @@ void intel_hdmi_probe(struct intel_encoder *intel_encoder)
 			enc_to_intel_hdmi(&intel_encoder->base);
 	struct intel_connector *intel_connector =
 				intel_hdmi->attached_connector;
+	struct drm_i915_private *dev_priv = to_i915(intel_encoder->base.dev);
+	bool live_status = false;
+	unsigned int retry = 3;
+
+	live_status = intel_digital_port_connected(dev_priv,
+						   hdmi_to_dig_port(intel_hdmi));
+	if (!intel_connector->detect_edid && live_status == false) {
+		/*
+		 * Hotplug had occurred and old status was disconnected,
+		 * so it might be possible that live status is not set,
+		 * so retry for few times
+		 */
+		do {
+			mdelay(10);
+			live_status = intel_digital_port_connected(dev_priv,
+						hdmi_to_dig_port(intel_hdmi));
+			if (live_status)
+				break;
+		} while (retry--);
+	}
 
 	/*
 	 * We are here, means there is a hotplug or a force
@@ -1381,7 +1402,7 @@ void intel_hdmi_probe(struct intel_encoder *intel_encoder)
 	 * DDC bus to check the current status of HDMI.
 	 */
 	intel_hdmi_unset_edid(&intel_connector->base);
-	if (intel_hdmi_set_edid(&intel_connector->base))
+	if (intel_hdmi_set_edid(&intel_connector->base, live_status))
 		DRM_DEBUG_DRIVER("DDC probe: got EDID\n");
 	else
 		DRM_DEBUG_DRIVER("DDC probe: no EDID\n");
@@ -1432,7 +1453,7 @@ intel_hdmi_force(struct drm_connector *connector)
 	if (connector->status != connector_status_connected)
 		return;
 
-	intel_hdmi_set_edid(connector);
+	intel_hdmi_set_edid(connector, true);
 	hdmi_to_dig_port(intel_hdmi)->base.type = INTEL_OUTPUT_HDMI;
 }
 
Signature · 6 lines
1.7.10.4

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

PATCH 1/1 drm/i915: Check live status before reading edid

The Bspec is very clear that Live status must be checked about before trying to read EDID over DDC channel. This patch makes sure that HDMI EDID is read only when live status us up.

The live status doesn't seem to perform very consistent across various platforms when tested with different monitors. The reason behind that is some monitors are late to provide right voltage to set live_status up. So, after getting the interrupt, for a small duration, live status reg fluctuates, and then settles down showing the correct staus.

This is explained here in, in a rough way:
HPD line ________________

			 |\ T1 = Monitor Hotplug causing IRQ
			 | \______________________________________
			 | |
                         | |
			 | |   T2 = Live status is stable
			 | |  _____________________________________
			 | | /|
Live status _____________|_|/ |
			 | |  |
			 | |  |
			 | |  |
			T0 T1  T2

(Between T1 and T2 Live status fluctuates or can be even low, depending on
the monitor)

After several experiments, we have concluded that a max delay of 30ms is enough to allow the live status to settle down with most of the monitors. This total delay of 30ms has been split into a resolution of 3 retries of 10ms each, for the better cases.

This delay is kept at 30ms, keeping in consideration that, HDCP compliance expect the HPD handler to respond a plug out in 100ms, by disabling port.

v2: Adding checks for VLV/CHV as well. Reusing old ibx and g4x functions to check digital port status. Adding a separate function to get bxt live status (Daniel)
v3: Using intel_encoder->hpd_pin to check the live status (Siva) Moving the live status read to intel_hdmi_probe and passing parameter to read/not to read the edid. (me)
v4:
* Added live status check for all platforms using intel_digital_port_connected.
* Rebased on top of Jani's DP cleanup series
* Some monitors take time in setting the live status. So retry for few times if this is a connect HPD
v5: Removed extra "drm/i915" from commit message. Adding Shashank's sob
which was missed.
v6: Drop the (!detect_edid && !live_status check) check because for DDI ports which are enumerated as hdmi as well as DP, we don't have a mechanism to differentiate between DP and hdmi inside the encoder's hot_plug. This leads to call to the hdmi's hot_plug hook for DP as well as hdmi which leads to issues during unplug because of the above check.

---
 drivers/gpu/drm/i915/intel_hdmi.c |   26 +++++++++++++++++++-------
 1 file changed, 19 insertions(+), 7 deletions(-)
drivers/gpu/drm/i915/intel_hdmi.c +19 -7
index 1eda71a..c6e9156 100644
@@ -1329,22 +1329,23 @@ intel_hdmi_unset_edid(struct drm_connector *connector)
 }
 
 static bool
-intel_hdmi_set_edid(struct drm_connector *connector)
+intel_hdmi_set_edid(struct drm_connector *connector, bool force)
 {
 	struct drm_i915_private *dev_priv = to_i915(connector->dev);
 	struct intel_hdmi *intel_hdmi = intel_attached_hdmi(connector);
 	struct intel_encoder *intel_encoder =
 		&hdmi_to_dig_port(intel_hdmi)->base;
 	enum intel_display_power_domain power_domain;
-	struct edid *edid;
+	struct edid *edid = NULL;
 	bool connected = false;
 
 	power_domain = intel_display_port_power_domain(intel_encoder);
 	intel_display_power_get(dev_priv, power_domain);
 
-	edid = drm_get_edid(connector,
-			    intel_gmbus_get_adapter(dev_priv,
-						    intel_hdmi->ddc_bus));
+	if (force)
+		edid = drm_get_edid(connector,
+				    intel_gmbus_get_adapter(dev_priv,
+				    intel_hdmi->ddc_bus));
 
 	intel_display_power_put(dev_priv, power_domain);
 
@@ -1374,6 +1375,17 @@ void intel_hdmi_probe(struct intel_encoder *intel_encoder)
 			enc_to_intel_hdmi(&intel_encoder->base);
 	struct intel_connector *intel_connector =
 				intel_hdmi->attached_connector;
+	struct drm_i915_private *dev_priv = to_i915(intel_encoder->base.dev);
+	bool live_status = false;
+	unsigned int retry = 3;
+
+	do {
+		live_status = intel_digital_port_connected(dev_priv,
+				hdmi_to_dig_port(intel_hdmi));
+		if (live_status)
+			break;
+		mdelay(10);
+	} while (retry--);
 
 	/*
 	 * We are here, means there is a hotplug or a force
@@ -1381,7 +1393,7 @@ void intel_hdmi_probe(struct intel_encoder *intel_encoder)
 	 * DDC bus to check the current status of HDMI.
 	 */
 	intel_hdmi_unset_edid(&intel_connector->base);
-	if (intel_hdmi_set_edid(&intel_connector->base))
+	if (intel_hdmi_set_edid(&intel_connector->base, live_status))
 		DRM_DEBUG_DRIVER("DDC probe: got EDID\n");
 	else
 		DRM_DEBUG_DRIVER("DDC probe: no EDID\n");
@@ -1432,7 +1444,7 @@ intel_hdmi_force(struct drm_connector *connector)
 	if (connector->status != connector_status_connected)
 		return;
 
-	intel_hdmi_set_edid(connector);
+	intel_hdmi_set_edid(connector, true);
 	hdmi_to_dig_port(intel_hdmi)->base.type = INTEL_OUTPUT_HDMI;
 }
 
Signature · 6 lines
1.7.10.4

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

Re: [PATCH 4/6] drm/i915: drm/i915: Check live status before reading edid

In reply to Daniel Vetter
Quoted 5 lines · Daniel Vetter
Quoted 3 lines · Sonika Jindal

The Bspec is very clear that Live status must be checked about before trying to read EDID over DDC channel. This patch makes sure that HDMI EDID is read only when live status us up.

s/us/is

Quoted 84 lines
Quoted 22 lines

The live status doesn't seem to perform very consistent across various platforms when tested with different monitors. The reason behind that is some monitors are late to provide right voltage to set live_status up. So, after getting the interrupt, for a small duration, live status reg fluctuates, and then settles down showing the correct staus.

This is explained here in, in a rough way:
HPD line ________________

                       |\ T1 = Monitor Hotplug causing IRQ
                       | \______________________________________
                       | |
                         | |
                       | |   T2 = Live status is stable
                       | |  _____________________________________
                       | | /|
Live status _____________|_|/ |
                       | |  |
                       | |  |
                       | |  |
                      T0 T1  T2
Quoted 1 line

(Between T1 and T2 Live status fluctuates or can be even low, depending

on

Quoted 8 lines

the monitor)

After several experiments, we have concluded that a max delay of 30ms is enough to allow the live status to settle down with most of the monitors. This total delay of 30ms has been split into a resolution of 3 retries of 10ms each, for the better cases.

This delay is kept at 30ms, keeping in consideration that, HDCP

compliance

Quoted 1 line

expect the HPD handler to respond a plug out in 100ms, by disabling the

port.

Quoted 14 lines

v2: Adding checks for VLV/CHV as well. Reusing old ibx and g4x functions to check digital port status. Adding a separate function to get bxt live status (Daniel)
v3: Using intel_encoder->hpd_pin to check the live status (Siva) Moving the live status read to intel_hdmi_probe and passing parameter to read/not to read the edid. (me)
v4:
* Added live status check for all platforms using intel_digital_port_connected.
* Rebased on top of Jani's DP cleanup series
* Some monitors take time in setting the live status. So retry for few times if this is a connect HPD

Quoted 1 line

Really pretty commit message! Had some minor comments on two other patches in this series while reading through them, but looks good otherwise. Please sign up someone for final review of the details and I'll pull in the patches once that's done.
-Daniel

Quoted 2 lines
---
 drivers/gpu/drm/i915/intel_hdmi.c |   35

++++++++++++++++++++++++++++-------

Review

Quoted 3 lines
 1 file changed, 28 insertions(+), 7 deletions(-)
(quoted diff) +0 -0

b/drivers/gpu/drm/i915/intel_hdmi.c

drivers/gpu/drm/i915/intel_hdmi.c intel_hdmi_unset_edid(struct drm_connector

Quoted 4 lines

index 1eda71a..d82887b 100644

drivers/gpu/drm/i915/intel_hdmi.c +0 -0
@@ -1329,22 +1329,23 @@ intel_hdmi_unset_edid(struct drm_connector

*connector)

Quoted 15 lines
drivers/gpu/drm/i915/intel_hdmi.c +2 -2
 }
 
 static bool
-intel_hdmi_set_edid(struct drm_connector *connector)
+intel_hdmi_set_edid(struct drm_connector *connector, bool force)
 {
      struct drm_i915_private *dev_priv = to_i915(connector->dev);
      struct intel_hdmi *intel_hdmi = intel_attached_hdmi(connector);
      struct intel_encoder *intel_encoder =
              &hdmi_to_dig_port(intel_hdmi)->base;
      enum intel_display_power_domain power_domain;
-     struct edid *edid;
+     struct edid *edid = NULL;
      bool connected = false;
 
Quoted 8 lines

drivers/gpu/drm/i915/intel_hdmi.c intel_hdmi_unset_edid(struct drm_connector *connector) · line 1345

Quoted 7 lines
drivers/gpu/drm/i915/intel_hdmi.c +1 -3
      power_domain = intel_display_port_power_domain(intel_encoder);
      intel_display_power_get(dev_priv, power_domain);
 
-     edid = drm_get_edid(connector,
-                         intel_gmbus_get_adapter(dev_priv,
-                                                 intel_hdmi->ddc_bus));
+     if (force)

The force name is confusing here since you only don't get edid in any other way when it is false...

But also there is the risk of this function to get called only with this false and never gets edid..
So I believe we should add a debug message here at least for this case so if we start getting cases where panel is taking more than 30ms to get live we will know why we did't actually read the edid....

Quoted 28 lines

drivers/gpu/drm/i915/intel_hdmi.c void intel_hdmi_probe(struct intel_encoder

Quoted 7 lines
drivers/gpu/drm/i915/intel_hdmi.c +3 -0
+             edid = drm_get_edid(connector,
+                                 intel_gmbus_get_adapter(dev_priv,
+                                 intel_hdmi->ddc_bus));
 
      intel_display_power_put(dev_priv, power_domain);
 
@@ -1374,6 +1375,26 @@ void intel_hdmi_probe(struct intel_encoder

*intel_encoder)

drivers/gpu/drm/i915/intel_hdmi.c void intel_hdmi_probe(struct intel_encoder *intel_encoder)

Quoted 4 lines
drivers/gpu/drm/i915/intel_hdmi.c +1 -0
                      enc_to_intel_hdmi(&intel_encoder->base);
      struct intel_connector *intel_connector =
                              intel_hdmi->attached_connector;
+     struct drm_i915_private *dev_priv =

to_i915(intel_encoder->base.dev);

drivers/gpu/drm/i915/intel_hdmi.c void intel_hdmi_probe(struct intel_encoder *intel_encoder)

Quoted 5 lines
drivers/gpu/drm/i915/intel_hdmi.c +5 -0
+     bool live_status = false;
+     unsigned int retry = 3;
+
+     live_status = intel_digital_port_connected(dev_priv,
+

hdmi_to_dig_port(intel_hdmi));

drivers/gpu/drm/i915/intel_hdmi.c void intel_hdmi_probe(struct intel_encoder *intel_encoder)

Quoted 9 lines
drivers/gpu/drm/i915/intel_hdmi.c +9 -0
+     if (!intel_connector->detect_edid && live_status == false) {
+             /*
+              * Hotplug had occurred and old status was disconnected,
+              * so it might be possible that live status is not set,
+              * so retry for few times
+              */
+             do {
+                     mdelay(10);
+                     live_status =
Quoted 7 lines

intel_digital_port_connected(dev_priv,

drivers/gpu/drm/i915/intel_hdmi.c void intel_hdmi_probe(struct intel_encoder *intel_encoder) · line 1381

Quoted 1 line
drivers/gpu/drm/i915/intel_hdmi.c +1 -0
+

hdmi_to_dig_port(intel_hdmi));

Quoted 3 lines
drivers/gpu/drm/i915/intel_hdmi.c +3 -0
+                     if (live_status)
+                             break;
+             } while (retry--);

small and optional bikesheding here: I'd do:
while(!live_status && --retry) {
mdelay(10);
live_status = ...
}

Quoted 26 lines

drivers/gpu/drm/i915/intel_hdmi.c void intel_hdmi_probe(struct intel_encoder

Quoted 5 lines
drivers/gpu/drm/i915/intel_hdmi.c +1 -0
+     }
 
      /*
       * We are here, means there is a hotplug or a force
@@ -1381,7 +1402,7 @@ void intel_hdmi_probe(struct intel_encoder

*intel_encoder)

Quoted 20 lines
drivers/gpu/drm/i915/intel_hdmi.c +2 -3
       * DDC bus to check the current status of HDMI.
       */
      intel_hdmi_unset_edid(&intel_connector->base);
-     if (intel_hdmi_set_edid(&intel_connector->base))
+     if (intel_hdmi_set_edid(&intel_connector->base, live_status))
              DRM_DEBUG_DRIVER("DDC probe: got EDID\n");
      else
              DRM_DEBUG_DRIVER("DDC probe: no EDID\n");
@@ -1432,7 +1453,7 @@ intel_hdmi_force(struct drm_connector *connector)
      if (connector->status != connector_status_connected)
              return;
 
-     intel_hdmi_set_edid(connector);
+     intel_hdmi_set_edid(connector, true);
      hdmi_to_dig_port(intel_hdmi)->base.type = INTEL_OUTPUT_HDMI;
 }
 
--

1.7.10.4

Quoted 14 lines
Quoted 4 lines

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

--
Daniel Vetter
Software Engineer, Intel Corporation
http://blog.ffwll.ch
_______________________________________________
Intel-gfx mailing list
[email protected]
http://lists.freedesktop.org/mailman/listinfo/intel-gfx

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

#45 lore raw

PATCH 1/1 drm/i915: Check live status before reading edid

The Bspec is very clear that Live status must be checked about before trying to read EDID over DDC channel. This patch makes sure that HDMI EDID is read only when live status is up.

The live status doesn't seem to perform very consistent across various platforms when tested with different monitors. The reason behind that is some monitors are late to provide right voltage to set live_status up. So, after getting the interrupt, for a small duration, live status reg fluctuates, and then settles down showing the correct staus.

This is explained here in, in a rough way:
HPD line ________________

			 |\ T1 = Monitor Hotplug causing IRQ
			 | \______________________________________
			 | |
                         | |
			 | |   T2 = Live status is stable
			 | |  _____________________________________
			 | | /|
Live status _____________|_|/ |
			 | |  |
			 | |  |
			 | |  |
			T0 T1  T2

(Between T1 and T2 Live status fluctuates or can be even low, depending on
the monitor)

After several experiments, we have concluded that a max delay of 30ms is enough to allow the live status to settle down with most of the monitors. This total delay of 30ms has been split into a resolution of 3 retries of 10ms each, for the better cases.

This delay is kept at 30ms, keeping in consideration that, HDCP compliance expect the HPD handler to respond a plug out in 100ms, by disabling port.

v2: Adding checks for VLV/CHV as well. Reusing old ibx and g4x functions to check digital port status. Adding a separate function to get bxt live status (Daniel)
v3: Using intel_encoder->hpd_pin to check the live status (Siva) Moving the live status read to intel_hdmi_probe and passing parameter to read/not to read the edid. (me)
v4:
* Added live status check for all platforms using intel_digital_port_connected.
* Rebased on top of Jani's DP cleanup series
* Some monitors take time in setting the live status. So retry for few times if this is a connect HPD
v5: Removed extra "drm/i915" from commit message. Adding Shashank's sob
which was missed.
v6: Drop the (!detect_edid && !live_status check) check because for DDI ports which are enumerated as hdmi as well as DP, we don't have a mechanism to differentiate between DP and hdmi inside the encoder's hot_plug. This leads to call to the hdmi's hot_plug hook for DP as well as hdmi which leads to issues during unplug because of the above check.
v7: Make intel_digital_port_connected global in this patch, some reformatting of while loop, adding a print when live status is not up. (Rodrigo)

---
 drivers/gpu/drm/i915/intel_dp.c   |    2 +-
 drivers/gpu/drm/i915/intel_drv.h  |    2 ++
 drivers/gpu/drm/i915/intel_hdmi.c |   26 +++++++++++++++++++-------
 3 files changed, 22 insertions(+), 8 deletions(-)
drivers/gpu/drm/i915/intel_dp.c +1 -1
index bf17030..fedf6d1 100644
@@ -4695,7 +4695,7 @@ static bool bxt_digital_port_connected(struct drm_i915_private *dev_priv,
  *
  * Return %true if @port is connected, %false otherwise.
  */
-static bool intel_digital_port_connected(struct drm_i915_private *dev_priv,
+bool intel_digital_port_connected(struct drm_i915_private *dev_priv,
 					 struct intel_digital_port *port)
 {
 	if (HAS_PCH_IBX(dev_priv))
drivers/gpu/drm/i915/intel_drv.h +2 -0
index b6c2c20..ac6d748 100644
@@ -1210,6 +1210,8 @@ void intel_edp_drrs_disable(struct intel_dp *intel_dp);
 void intel_edp_drrs_invalidate(struct drm_device *dev,
 		unsigned frontbuffer_bits);
 void intel_edp_drrs_flush(struct drm_device *dev, unsigned frontbuffer_bits);
+bool intel_digital_port_connected(struct drm_i915_private *dev_priv,
+					 struct intel_digital_port *port);
 
 /* intel_dp_mst.c */
 int intel_dp_mst_encoder_init(struct intel_digital_port *intel_dig_port, int conn_id);
drivers/gpu/drm/i915/intel_hdmi.c +19 -7
index 1eda71a..d366ca5 100644
@@ -1329,22 +1329,23 @@ intel_hdmi_unset_edid(struct drm_connector *connector)
 }
 
 static bool
-intel_hdmi_set_edid(struct drm_connector *connector)
+intel_hdmi_set_edid(struct drm_connector *connector, bool force)
 {
 	struct drm_i915_private *dev_priv = to_i915(connector->dev);
 	struct intel_hdmi *intel_hdmi = intel_attached_hdmi(connector);
 	struct intel_encoder *intel_encoder =
 		&hdmi_to_dig_port(intel_hdmi)->base;
 	enum intel_display_power_domain power_domain;
-	struct edid *edid;
+	struct edid *edid = NULL;
 	bool connected = false;
 
 	power_domain = intel_display_port_power_domain(intel_encoder);
 	intel_display_power_get(dev_priv, power_domain);
 
-	edid = drm_get_edid(connector,
-			    intel_gmbus_get_adapter(dev_priv,
-						    intel_hdmi->ddc_bus));
+	if (force)
+		edid = drm_get_edid(connector,
+				    intel_gmbus_get_adapter(dev_priv,
+				    intel_hdmi->ddc_bus));
 
 	intel_display_power_put(dev_priv, power_domain);
 
@@ -1374,14 +1375,25 @@ void intel_hdmi_probe(struct intel_encoder *intel_encoder)
 			enc_to_intel_hdmi(&intel_encoder->base);
 	struct intel_connector *intel_connector =
 				intel_hdmi->attached_connector;
+	struct drm_i915_private *dev_priv = to_i915(intel_encoder->base.dev);
+	bool live_status = false;
+	unsigned int retry = 3;
+
+	while (!live_status && --retry) {
+		live_status = intel_digital_port_connected(dev_priv,
+				hdmi_to_dig_port(intel_hdmi));
+		mdelay(10);
+	}
 
+	if (!live_status)
+		DRM_DEBUG_KMS("Live status not up!");
 	/*
 	 * We are here, means there is a hotplug or a force
 	 * detection. Clear the cached EDID and probe the
 	 * DDC bus to check the current status of HDMI.
 	 */
 	intel_hdmi_unset_edid(&intel_connector->base);
-	if (intel_hdmi_set_edid(&intel_connector->base))
+	if (intel_hdmi_set_edid(&intel_connector->base, live_status))
 		DRM_DEBUG_DRIVER("DDC probe: got EDID\n");
 	else
 		DRM_DEBUG_DRIVER("DDC probe: no EDID\n");
@@ -1432,7 +1444,7 @@ intel_hdmi_force(struct drm_connector *connector)
 	if (connector->status != connector_status_connected)
 		return;
 
-	intel_hdmi_set_edid(connector);
+	intel_hdmi_set_edid(connector, true);
 	hdmi_to_dig_port(intel_hdmi)->base.type = INTEL_OUTPUT_HDMI;
 }
 
Signature · 6 lines
1.7.10.4

_______________________________________________
Intel-gfx mailing list
[email protected]
http://lists.freedesktop.org/mailman/listinfo/intel-gfx
Reviewed-by on v1 1/1
#46 lore raw

Re: [PATCH] drm/i915: Check live status before reading edid

Thanks

Quoted 176 lines · Sonika Jindal

The Bspec is very clear that Live status must be checked about before trying to read EDID over DDC channel. This patch makes sure that HDMI EDID is read only when live status is up.

The live status doesn't seem to perform very consistent across various platforms when tested with different monitors. The reason behind that is some monitors are late to provide right voltage to set live_status up. So, after getting the interrupt, for a small duration, live status reg fluctuates, and then settles down showing the correct staus.

This is explained here in, in a rough way:
HPD line ________________

                         |\ T1 = Monitor Hotplug causing IRQ
                         | \______________________________________
                         | |
                         | |
                         | |   T2 = Live status is stable
                         | |  _____________________________________
                         | | /|
Live status _____________|_|/ |
                         | |  |
                         | |  |
                         | |  |
                        T0 T1  T2

(Between T1 and T2 Live status fluctuates or can be even low, depending on
the monitor)

After several experiments, we have concluded that a max delay of 30ms is enough to allow the live status to settle down with most of the monitors. This total delay of 30ms has been split into a resolution of 3 retries of 10ms each, for the better cases.

This delay is kept at 30ms, keeping in consideration that, HDCP compliance expect the HPD handler to respond a plug out in 100ms, by disabling port.

v2: Adding checks for VLV/CHV as well. Reusing old ibx and g4x functions to check digital port status. Adding a separate function to get bxt live status (Daniel)
v3: Using intel_encoder->hpd_pin to check the live status (Siva) Moving the live status read to intel_hdmi_probe and passing parameter to read/not to read the edid. (me)
v4:
* Added live status check for all platforms using intel_digital_port_connected.
* Rebased on top of Jani's DP cleanup series
* Some monitors take time in setting the live status. So retry for few times if this is a connect HPD
v5: Removed extra "drm/i915" from commit message. Adding Shashank's sob
which was missed.
v6: Drop the (!detect_edid && !live_status check) check because for DDI ports which are enumerated as hdmi as well as DP, we don't have a mechanism to differentiate between DP and hdmi inside the encoder's hot_plug. This leads to call to the hdmi's hot_plug hook for DP as well as hdmi which leads to issues during unplug because of the above check.
v7: Make intel_digital_port_connected global in this patch, some reformatting of while loop, adding a print when live status is not up. (Rodrigo)

---
 drivers/gpu/drm/i915/intel_dp.c   |    2 +-
 drivers/gpu/drm/i915/intel_drv.h  |    2 ++
 drivers/gpu/drm/i915/intel_hdmi.c |   26 +++++++++++++++++++-------
 3 files changed, 22 insertions(+), 8 deletions(-)
drivers/gpu/drm/i915/intel_dp.c +1 -1
index bf17030..fedf6d1 100644
@@ -4695,7 +4695,7 @@ static bool bxt_digital_port_connected(struct
 drm_i915_private *dev_priv,
  *
  * Return %true if @port is connected, %false otherwise.
  */
-static bool intel_digital_port_connected(struct drm_i915_private
 *dev_priv,
+bool intel_digital_port_connected(struct drm_i915_private *dev_priv,
                                         struct intel_digital_port *port)
 {
        if (HAS_PCH_IBX(dev_priv))
drivers/gpu/drm/i915/intel_drv.h +2 -0
index b6c2c20..ac6d748 100644
@@ -1210,6 +1210,8 @@ void intel_edp_drrs_disable(struct intel_dp
 *intel_dp);
 void intel_edp_drrs_invalidate(struct drm_device *dev,
                unsigned frontbuffer_bits);
 void intel_edp_drrs_flush(struct drm_device *dev, unsigned
 frontbuffer_bits);
+bool intel_digital_port_connected(struct drm_i915_private *dev_priv,
+                                        struct intel_digital_port *port);
 

/* intel_dp_mst.c */
int intel_dp_mst_encoder_init(struct intel_digital_port *intel_dig_port,
int conn_id);

drivers/gpu/drm/i915/intel_hdmi.c +18 -6
index 1eda71a..d366ca5 100644
@@ -1329,22 +1329,23 @@ intel_hdmi_unset_edid(struct drm_connector
 *connector)
 }
 
 static bool
-intel_hdmi_set_edid(struct drm_connector *connector)
+intel_hdmi_set_edid(struct drm_connector *connector, bool force)
 {
        struct drm_i915_private *dev_priv = to_i915(connector->dev);
        struct intel_hdmi *intel_hdmi = intel_attached_hdmi(connector);
        struct intel_encoder *intel_encoder =
                &hdmi_to_dig_port(intel_hdmi)->base;
        enum intel_display_power_domain power_domain;
-       struct edid *edid;
+       struct edid *edid = NULL;
        bool connected = false;
 
        power_domain = intel_display_port_power_domain(intel_encoder);
        intel_display_power_get(dev_priv, power_domain);
 
-       edid = drm_get_edid(connector,
-                           intel_gmbus_get_adapter(dev_priv,
-                                                   intel_hdmi->ddc_bus));
+       if (force)
+               edid = drm_get_edid(connector,
+                                   intel_gmbus_get_adapter(dev_priv,
+                                   intel_hdmi->ddc_bus));
 
        intel_display_power_put(dev_priv, power_domain);
@@ -1374,14 +1375,25 @@ void intel_hdmi_probe(struct intel_encoder
 *intel_encoder)
                        enc_to_intel_hdmi(&intel_encoder->base);
        struct intel_connector *intel_connector =
                                intel_hdmi->attached_connector;
+       struct drm_i915_private *dev_priv =
 to_i915(intel_encoder->base.dev);
+       bool live_status = false;
+       unsigned int retry = 3;
+
+       while (!live_status && --retry) {
+               live_status = intel_digital_port_connected(dev_priv,
+                               hdmi_to_dig_port(intel_hdmi));
+               mdelay(10);
+       }
 
+       if (!live_status)
+               DRM_DEBUG_KMS("Live status not up!");
        /*
         * We are here, means there is a hotplug or a force
         * detection. Clear the cached EDID and probe the
         * DDC bus to check the current status of HDMI.
         */
        intel_hdmi_unset_edid(&intel_connector->base);
-       if (intel_hdmi_set_edid(&intel_connector->base))
+       if (intel_hdmi_set_edid(&intel_connector->base, live_status))
                DRM_DEBUG_DRIVER("DDC probe: got EDID\n");
        else
                DRM_DEBUG_DRIVER("DDC probe: no EDID\n");
drivers/gpu/drm/i915/intel_hdmi.c +1 -2
@@ -1432,7 +1444,7 @@ intel_hdmi_force(struct drm_connector *connector)
        if (connector->status != connector_status_connected)
                return;
 
-       intel_hdmi_set_edid(connector);
+       intel_hdmi_set_edid(connector, true);
        hdmi_to_dig_port(intel_hdmi)->base.type = INTEL_OUTPUT_HDMI;
 }
 
--

1.7.10.4

Quoted 5 lines

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

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

#47 lore raw

Re: [PATCH] drm/i915: Check live status before reading edid

Quoted 69 lines · Rodrigo Vivi

Thanks

Quoted 61 lines · Sonika Jindal

The Bspec is very clear that Live status must be checked about before trying to read EDID over DDC channel. This patch makes sure that HDMI EDID is read only when live status is up.

The live status doesn't seem to perform very consistent across various platforms when tested with different monitors. The reason behind that is some monitors are late to provide right voltage to set live_status up. So, after getting the interrupt, for a small duration, live status reg fluctuates, and then settles down showing the correct staus.

This is explained here in, in a rough way:
HPD line ________________

                         |\ T1 = Monitor Hotplug causing IRQ
                         | \______________________________________
                         | |
                         | |
                         | |   T2 = Live status is stable
                         | |  _____________________________________
                         | | /|
Live status _____________|_|/ |
                         | |  |
                         | |  |
                         | |  |
                        T0 T1  T2

(Between T1 and T2 Live status fluctuates or can be even low, depending on
the monitor)

After several experiments, we have concluded that a max delay of 30ms is enough to allow the live status to settle down with most of the monitors. This total delay of 30ms has been split into a resolution of 3 retries of 10ms each, for the better cases.

This delay is kept at 30ms, keeping in consideration that, HDCP compliance expect the HPD handler to respond a plug out in 100ms, by disabling port.

v2: Adding checks for VLV/CHV as well. Reusing old ibx and g4x functions to check digital port status. Adding a separate function to get bxt live status (Daniel)
v3: Using intel_encoder->hpd_pin to check the live status (Siva) Moving the live status read to intel_hdmi_probe and passing parameter to read/not to read the edid. (me)
v4:
* Added live status check for all platforms using intel_digital_port_connected.
* Rebased on top of Jani's DP cleanup series
* Some monitors take time in setting the live status. So retry for few times if this is a connect HPD
v5: Removed extra "drm/i915" from commit message. Adding Shashank's sob
which was missed.
v6: Drop the (!detect_edid && !live_status check) check because for DDI ports which are enumerated as hdmi as well as DP, we don't have a mechanism to differentiate between DP and hdmi inside the encoder's hot_plug. This leads to call to the hdmi's hot_plug hook for DP as well as hdmi which leads to issues during unplug because of the above check.
v7: Make intel_digital_port_connected global in this patch, some reformatting of while loop, adding a print when live status is not up. (Rodrigo)

Since this is tricky stuff and other patches in this series are blocked until we have a clearer picture can you please rebase this to be the first patch on top of -nightly so that I can pull it in directly?

I tried to do that but proved a bit too messy.
-Daniel

Quoted 125 lines
Quoted 28 lines
---
 drivers/gpu/drm/i915/intel_dp.c   |    2 +-
 drivers/gpu/drm/i915/intel_drv.h  |    2 ++
 drivers/gpu/drm/i915/intel_hdmi.c |   26 +++++++++++++++++++-------
 3 files changed, 22 insertions(+), 8 deletions(-)
drivers/gpu/drm/i915/intel_dp.c +1 -1
index bf17030..fedf6d1 100644
@@ -4695,7 +4695,7 @@ static bool bxt_digital_port_connected(struct
 drm_i915_private *dev_priv,
  *
  * Return %true if @port is connected, %false otherwise.
  */
-static bool intel_digital_port_connected(struct drm_i915_private
 *dev_priv,
+bool intel_digital_port_connected(struct drm_i915_private *dev_priv,
                                         struct intel_digital_port *port)
 {
        if (HAS_PCH_IBX(dev_priv))
drivers/gpu/drm/i915/intel_drv.h +0 -0
index b6c2c20..ac6d748 100644

drivers/gpu/drm/i915/intel_drv.h void intel_edp_drrs_disable(struct intel_dp · line 1216

Quoted 7 lines
drivers/gpu/drm/i915/intel_drv.h +2 -0
@@ -1210,6 +1210,8 @@ void intel_edp_drrs_disable(struct intel_dp
 *intel_dp);
 void intel_edp_drrs_invalidate(struct drm_device *dev,
                unsigned frontbuffer_bits);
 void intel_edp_drrs_flush(struct drm_device *dev, unsigned
 frontbuffer_bits);
+bool intel_digital_port_connected(struct drm_i915_private *dev_priv,
+                                        struct intel_digital_port *port);

g>

Quoted 79 lines
drivers/gpu/drm/i915/intel_drv.h +0 -0
 /* intel_dp_mst.c */

int intel_dp_mst_encoder_init(struct intel_digital_port *intel_dig_port,
int conn_id);

drivers/gpu/drm/i915/intel_hdmi.c +18 -6
index 1eda71a..d366ca5 100644
@@ -1329,22 +1329,23 @@ intel_hdmi_unset_edid(struct drm_connector
 *connector)
 }
 
 static bool
-intel_hdmi_set_edid(struct drm_connector *connector)
+intel_hdmi_set_edid(struct drm_connector *connector, bool force)
 {
        struct drm_i915_private *dev_priv = to_i915(connector->dev);
        struct intel_hdmi *intel_hdmi = intel_attached_hdmi(connector);
        struct intel_encoder *intel_encoder =
                &hdmi_to_dig_port(intel_hdmi)->base;
        enum intel_display_power_domain power_domain;
-       struct edid *edid;
+       struct edid *edid = NULL;
        bool connected = false;
 
        power_domain = intel_display_port_power_domain(intel_encoder);
        intel_display_power_get(dev_priv, power_domain);
 
-       edid = drm_get_edid(connector,
-                           intel_gmbus_get_adapter(dev_priv,
-                                                   intel_hdmi->ddc_bus));
+       if (force)
+               edid = drm_get_edid(connector,
+                                   intel_gmbus_get_adapter(dev_priv,
+                                   intel_hdmi->ddc_bus));
 
        intel_display_power_put(dev_priv, power_domain);
@@ -1374,14 +1375,25 @@ void intel_hdmi_probe(struct intel_encoder
 *intel_encoder)
                        enc_to_intel_hdmi(&intel_encoder->base);
        struct intel_connector *intel_connector =
                                intel_hdmi->attached_connector;
+       struct drm_i915_private *dev_priv =
 to_i915(intel_encoder->base.dev);
+       bool live_status = false;
+       unsigned int retry = 3;
+
+       while (!live_status && --retry) {
+               live_status = intel_digital_port_connected(dev_priv,
+                               hdmi_to_dig_port(intel_hdmi));
+               mdelay(10);
+       }
 
+       if (!live_status)
+               DRM_DEBUG_KMS("Live status not up!");
        /*
         * We are here, means there is a hotplug or a force
         * detection. Clear the cached EDID and probe the
         * DDC bus to check the current status of HDMI.
         */
        intel_hdmi_unset_edid(&intel_connector->base);
-       if (intel_hdmi_set_edid(&intel_connector->base))
+       if (intel_hdmi_set_edid(&intel_connector->base, live_status))
                DRM_DEBUG_DRIVER("DDC probe: got EDID\n");
        else
                DRM_DEBUG_DRIVER("DDC probe: no EDID\n");
drivers/gpu/drm/i915/intel_hdmi.c +1 -2
@@ -1432,7 +1444,7 @@ intel_hdmi_force(struct drm_connector *connector)
        if (connector->status != connector_status_connected)
                return;
 
-       intel_hdmi_set_edid(connector);
+       intel_hdmi_set_edid(connector, true);
        hdmi_to_dig_port(intel_hdmi)->base.type = INTEL_OUTPUT_HDMI;
 }
 
--

1.7.10.4

Quoted 5 lines

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

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

Signature · 7 lines
Daniel Vetter
Software Engineer, Intel Corporation
http://blog.ffwll.ch
_______________________________________________
Intel-gfx mailing list
[email protected]
http://lists.freedesktop.org/mailman/listinfo/intel-gfx
#48 lore raw

Re: [PATCH] drm/i915: Check live status before reading edid

Quoted 71 lines · Daniel Vetter
Quoted 69 lines · Rodrigo Vivi

Thanks

Quoted 61 lines · Sonika Jindal

The Bspec is very clear that Live status must be checked about before trying to read EDID over DDC channel. This patch makes sure that HDMI EDID is read only when live status is up.

The live status doesn't seem to perform very consistent across various platforms when tested with different monitors. The reason behind that is some monitors are late to provide right voltage to set live_status up. So, after getting the interrupt, for a small duration, live status reg fluctuates, and then settles down showing the correct staus.

This is explained here in, in a rough way:
HPD line ________________

                          |\ T1 = Monitor Hotplug causing IRQ
                          | \______________________________________
                          | |
                          | |
                          | |   T2 = Live status is stable
                          | |  _____________________________________
                          | | /|
Live status _____________|_|/ |
                          | |  |
                          | |  |
                          | |  |
                         T0 T1  T2

(Between T1 and T2 Live status fluctuates or can be even low, depending on
the monitor)

After several experiments, we have concluded that a max delay of 30ms is enough to allow the live status to settle down with most of the monitors. This total delay of 30ms has been split into a resolution of 3 retries of 10ms each, for the better cases.

This delay is kept at 30ms, keeping in consideration that, HDCP compliance expect the HPD handler to respond a plug out in 100ms, by disabling port.

v2: Adding checks for VLV/CHV as well. Reusing old ibx and g4x functions to check digital port status. Adding a separate function to get bxt live status (Daniel)
v3: Using intel_encoder->hpd_pin to check the live status (Siva) Moving the live status read to intel_hdmi_probe and passing parameter to read/not to read the edid. (me)
v4:
* Added live status check for all platforms using intel_digital_port_connected.
* Rebased on top of Jani's DP cleanup series
* Some monitors take time in setting the live status. So retry for few times if this is a connect HPD
v5: Removed extra "drm/i915" from commit message. Adding Shashank's sob
which was missed.
v6: Drop the (!detect_edid && !live_status check) check because for DDI ports which are enumerated as hdmi as well as DP, we don't have a mechanism to differentiate between DP and hdmi inside the encoder's hot_plug. This leads to call to the hdmi's hot_plug hook for DP as well as hdmi which leads to issues during unplug because of the above check.
v7: Make intel_digital_port_connected global in this patch, some reformatting of while loop, adding a print when live status is not up. (Rodrigo)

Quoted 7 lines

Since this is tricky stuff and other patches in this series are blocked until we have a clearer picture can you please rebase this to be the first patch on top of -nightly so that I can pull it in directly?

I tried to do that but proved a bit too messy.
-Daniel

Hmm, since rebasing this will require these changes (check for live status) to go in detect and then later when we reach a conclusion on hot_plug hook, we would need to move it to that function.
I think lets wait for the conclusion on the placement of that function.
What do you suggest?

Regards,
Sonika

Quoted 115 lines
Quoted 115 lines
Quoted 28 lines
---
  drivers/gpu/drm/i915/intel_dp.c   |    2 +-
  drivers/gpu/drm/i915/intel_drv.h  |    2 ++
  drivers/gpu/drm/i915/intel_hdmi.c |   26 +++++++++++++++++++-------
  3 files changed, 22 insertions(+), 8 deletions(-)
drivers/gpu/drm/i915/intel_dp.c +1 -1
index bf17030..fedf6d1 100644
@@ -4695,7 +4695,7 @@ static bool bxt_digital_port_connected(struct
 drm_i915_private *dev_priv,
   *
   * Return %true if @port is connected, %false otherwise.
   */
-static bool intel_digital_port_connected(struct drm_i915_private
 *dev_priv,
+bool intel_digital_port_connected(struct drm_i915_private *dev_priv,
                                          struct intel_digital_port *port)
  {
         if (HAS_PCH_IBX(dev_priv))
drivers/gpu/drm/i915/intel_drv.h +0 -0
index b6c2c20..ac6d748 100644

drivers/gpu/drm/i915/intel_drv.h void intel_edp_drrs_disable(struct intel_dp · line 1216

Quoted 7 lines
drivers/gpu/drm/i915/intel_drv.h +2 -0
@@ -1210,6 +1210,8 @@ void intel_edp_drrs_disable(struct intel_dp
 *intel_dp);
  void intel_edp_drrs_invalidate(struct drm_device *dev,
                 unsigned frontbuffer_bits);
  void intel_edp_drrs_flush(struct drm_device *dev, unsigned
 frontbuffer_bits);
+bool intel_digital_port_connected(struct drm_i915_private *dev_priv,
+                                        struct intel_digital_port *port);

g>

Quoted 79 lines
drivers/gpu/drm/i915/intel_drv.h +0 -0
  /* intel_dp_mst.c */

int intel_dp_mst_encoder_init(struct intel_digital_port *intel_dig_port,
int conn_id);

drivers/gpu/drm/i915/intel_hdmi.c +18 -6
index 1eda71a..d366ca5 100644
@@ -1329,22 +1329,23 @@ intel_hdmi_unset_edid(struct drm_connector
 *connector)
  }
 
  static bool
-intel_hdmi_set_edid(struct drm_connector *connector)
+intel_hdmi_set_edid(struct drm_connector *connector, bool force)
  {
         struct drm_i915_private *dev_priv = to_i915(connector->dev);
         struct intel_hdmi *intel_hdmi = intel_attached_hdmi(connector);
         struct intel_encoder *intel_encoder =
                 &hdmi_to_dig_port(intel_hdmi)->base;
         enum intel_display_power_domain power_domain;
-       struct edid *edid;
+       struct edid *edid = NULL;
         bool connected = false;
 
         power_domain = intel_display_port_power_domain(intel_encoder);
         intel_display_power_get(dev_priv, power_domain);
 
-       edid = drm_get_edid(connector,
-                           intel_gmbus_get_adapter(dev_priv,
-                                                   intel_hdmi->ddc_bus));
+       if (force)
+               edid = drm_get_edid(connector,
+                                   intel_gmbus_get_adapter(dev_priv,
+                                   intel_hdmi->ddc_bus));
 
         intel_display_power_put(dev_priv, power_domain);
@@ -1374,14 +1375,25 @@ void intel_hdmi_probe(struct intel_encoder
 *intel_encoder)
                         enc_to_intel_hdmi(&intel_encoder->base);
         struct intel_connector *intel_connector =
                                 intel_hdmi->attached_connector;
+       struct drm_i915_private *dev_priv =
 to_i915(intel_encoder->base.dev);
+       bool live_status = false;
+       unsigned int retry = 3;
+
+       while (!live_status && --retry) {
+               live_status = intel_digital_port_connected(dev_priv,
+                               hdmi_to_dig_port(intel_hdmi));
+               mdelay(10);
+       }
 
+       if (!live_status)
+               DRM_DEBUG_KMS("Live status not up!");
         /*
          * We are here, means there is a hotplug or a force
          * detection. Clear the cached EDID and probe the
          * DDC bus to check the current status of HDMI.
          */
         intel_hdmi_unset_edid(&intel_connector->base);
-       if (intel_hdmi_set_edid(&intel_connector->base))
+       if (intel_hdmi_set_edid(&intel_connector->base, live_status))
                 DRM_DEBUG_DRIVER("DDC probe: got EDID\n");
         else
                 DRM_DEBUG_DRIVER("DDC probe: no EDID\n");
drivers/gpu/drm/i915/intel_hdmi.c +1 -2
@@ -1432,7 +1444,7 @@ intel_hdmi_force(struct drm_connector *connector)
         if (connector->status != connector_status_connected)
                 return;
 
-       intel_hdmi_set_edid(connector);
+       intel_hdmi_set_edid(connector, true);
         hdmi_to_dig_port(intel_hdmi)->base.type = INTEL_OUTPUT_HDMI;
  }
 
--

1.7.10.4

Quoted 12 lines
Quoted 10 lines
Quoted 5 lines

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

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

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

#49 lore raw

Re: [PATCH] drm/i915: Check live status before reading edid

Quoted 78 lines · Sonika
Quoted 75 lines · Daniel Vetter
Quoted 69 lines · Rodrigo Vivi

Thanks

Quoted 61 lines · Sonika Jindal

The Bspec is very clear that Live status must be checked about before trying to read EDID over DDC channel. This patch makes sure that HDMI EDID is read only when live status is up.

The live status doesn't seem to perform very consistent across various platforms when tested with different monitors. The reason behind that is some monitors are late to provide right voltage to set live_status up. So, after getting the interrupt, for a small duration, live status reg fluctuates, and then settles down showing the correct staus.

This is explained here in, in a rough way:
HPD line ________________

                         |\ T1 = Monitor Hotplug causing IRQ
                         | \______________________________________
                         | |
                         | |
                         | |   T2 = Live status is stable
                         | |  _____________________________________
                         | | /|
Live status _____________|_|/ |
                         | |  |
                         | |  |
                         | |  |
                        T0 T1  T2

(Between T1 and T2 Live status fluctuates or can be even low, depending on
the monitor)

After several experiments, we have concluded that a max delay of 30ms is enough to allow the live status to settle down with most of the monitors. This total delay of 30ms has been split into a resolution of 3 retries of 10ms each, for the better cases.

This delay is kept at 30ms, keeping in consideration that, HDCP compliance expect the HPD handler to respond a plug out in 100ms, by disabling port.

v2: Adding checks for VLV/CHV as well. Reusing old ibx and g4x functions to check digital port status. Adding a separate function to get bxt live status (Daniel)
v3: Using intel_encoder->hpd_pin to check the live status (Siva) Moving the live status read to intel_hdmi_probe and passing parameter to read/not to read the edid. (me)
v4:
* Added live status check for all platforms using intel_digital_port_connected.
* Rebased on top of Jani's DP cleanup series
* Some monitors take time in setting the live status. So retry for few times if this is a connect HPD
v5: Removed extra "drm/i915" from commit message. Adding Shashank's sob
which was missed.
v6: Drop the (!detect_edid && !live_status check) check because for DDI ports which are enumerated as hdmi as well as DP, we don't have a mechanism to differentiate between DP and hdmi inside the encoder's hot_plug. This leads to call to the hdmi's hot_plug hook for DP as well as hdmi which leads to issues during unplug because of the above check.
v7: Make intel_digital_port_connected global in this patch, some reformatting of while loop, adding a print when live status is not up. (Rodrigo)

Since this is tricky stuff and other patches in this series are blocked until we have a clearer picture can you please rebase this to be the first patch on top of -nightly so that I can pull it in directly?

Quoted 8 lines
Quoted 3 lines

I tried to do that but proved a bit too messy.
-Daniel

Hmm, since rebasing this will require these changes (check for live status) to go in detect and then later when we reach a conclusion on hot_plug hook, we would need to move it to that function.
I think lets wait for the conclusion on the placement of that function. What do you suggest?

I think given how tricky this patch here is with testing it's preferably to get it int as soon as possible. Yes that means shuffling things around twice. But since the structure of how we'll handle hpd isn't clear yet at all there's a really high chance that you need to rework this patch anyway. So I don't expect doing this rebase right now will be more work. And it will give us better testing coverage, which is always good to have.

Thanks, Daniel

Signature · 7 lines
Daniel Vetter
Software Engineer, Intel Corporation
http://blog.ffwll.ch
_______________________________________________
Intel-gfx mailing list
[email protected]
http://lists.freedesktop.org/mailman/listinfo/intel-gfx
#50 lore raw

PATCH 1/1 drm/i915: Check live status before reading edid

The Bspec is very clear that Live status must be checked about before trying to read EDID over DDC channel. This patch makes sure that HDMI EDID is read only when live status is up.

The live status doesn't seem to perform very consistent across various platforms when tested with different monitors. The reason behind that is some monitors are late to provide right voltage to set live_status up. So, after getting the interrupt, for a small duration, live status reg fluctuates, and then settles down showing the correct staus.

This is explained here in, in a rough way:
HPD line ________________

			 |\ T1 = Monitor Hotplug causing IRQ
			 | \______________________________________
			 | |
                         | |
			 | |   T2 = Live status is stable
			 | |  _____________________________________
			 | | /|
Live status _____________|_|/ |
			 | |  |
			 | |  |
			 | |  |
			T0 T1  T2

(Between T1 and T2 Live status fluctuates or can be even low, depending on
the monitor)

After several experiments, we have concluded that a max delay of 30ms is enough to allow the live status to settle down with most of the monitors. This total delay of 30ms has been split into a resolution of 3 retries of 10ms each, for the better cases.

This delay is kept at 30ms, keeping in consideration that, HDCP compliance expect the HPD handler to respond a plug out in 100ms, by disabling port.

v2: Adding checks for VLV/CHV as well. Reusing old ibx and g4x functions to check digital port status. Adding a separate function to get bxt live status (Daniel)
v3: Using intel_encoder->hpd_pin to check the live status (Siva) Moving the live status read to intel_hdmi_probe and passing parameter to read/not to read the edid. (me)
v4:
* Added live status check for all platforms using intel_digital_port_connected.
* Rebased on top of Jani's DP cleanup series
* Some monitors take time in setting the live status. So retry for few times if this is a connect HPD
v5: Removed extra "drm/i915" from commit message. Adding Shashank's sob
which was missed.
v6: Drop the (!detect_edid && !live_status check) check because for DDI ports which are enumerated as hdmi as well as DP, we don't have a mechanism to differentiate between DP and hdmi inside the encoder's hot_plug. This leads to call to the hdmi's hot_plug hook for DP as well as hdmi which leads to issues during unplug because of the above check.
v7: Make intel_digital_port_connected global in this patch, some reformatting of while loop, adding a print when live status is not up. (Rodrigo)
v8: Rebase it on nightly which involved skipping the hot_plug hook for now and letting the live_status check happen in detect until the hpd handling part is finalized (Daniel)

---
 drivers/gpu/drm/i915/intel_dp.c   |    2 +-
 drivers/gpu/drm/i915/intel_drv.h  |    2 ++
 drivers/gpu/drm/i915/intel_hdmi.c |   28 +++++++++++++++++++++-------
 3 files changed, 24 insertions(+), 8 deletions(-)
drivers/gpu/drm/i915/intel_dp.c +1 -1
index a687250..a6e22dd 100644
@@ -4670,7 +4670,7 @@ static bool bxt_digital_port_connected(struct drm_i915_private *dev_priv,
  *
  * Return %true if @port is connected, %false otherwise.
  */
-static bool intel_digital_port_connected(struct drm_i915_private *dev_priv,
+bool intel_digital_port_connected(struct drm_i915_private *dev_priv,
 					 struct intel_digital_port *port)
 {
 	if (HAS_PCH_IBX(dev_priv))
drivers/gpu/drm/i915/intel_drv.h +2 -0
index 08ba362..1e01350 100644
@@ -1210,6 +1210,8 @@ void intel_edp_drrs_disable(struct intel_dp *intel_dp);
 void intel_edp_drrs_invalidate(struct drm_device *dev,
 		unsigned frontbuffer_bits);
 void intel_edp_drrs_flush(struct drm_device *dev, unsigned frontbuffer_bits);
+bool intel_digital_port_connected(struct drm_i915_private *dev_priv,
+					 struct intel_digital_port *port);
 void hsw_dp_set_ddi_pll_sel(struct intel_crtc_state *pipe_config);
 
 /* intel_dp_mst.c */
drivers/gpu/drm/i915/intel_hdmi.c +21 -7
index e978c59..bb33c66 100644
@@ -1329,22 +1329,23 @@ intel_hdmi_unset_edid(struct drm_connector *connector)
 }
 
 static bool
-intel_hdmi_set_edid(struct drm_connector *connector)
+intel_hdmi_set_edid(struct drm_connector *connector, bool force)
 {
 	struct drm_i915_private *dev_priv = to_i915(connector->dev);
 	struct intel_hdmi *intel_hdmi = intel_attached_hdmi(connector);
 	struct intel_encoder *intel_encoder =
 		&hdmi_to_dig_port(intel_hdmi)->base;
 	enum intel_display_power_domain power_domain;
-	struct edid *edid;
+	struct edid *edid = NULL;
 	bool connected = false;
 
 	power_domain = intel_display_port_power_domain(intel_encoder);
 	intel_display_power_get(dev_priv, power_domain);
 
-	edid = drm_get_edid(connector,
-			    intel_gmbus_get_adapter(dev_priv,
-						    intel_hdmi->ddc_bus));
+	if (force)
+		edid = drm_get_edid(connector,
+				    intel_gmbus_get_adapter(dev_priv,
+				    intel_hdmi->ddc_bus));
 
 	intel_display_power_put(dev_priv, power_domain);
 
@@ -1372,13 +1373,26 @@ static enum drm_connector_status
 intel_hdmi_detect(struct drm_connector *connector, bool force)
 {
 	enum drm_connector_status status;
+	struct intel_hdmi *intel_hdmi = intel_attached_hdmi(connector);
+	struct drm_i915_private *dev_priv = to_i915(connector->dev);
+	bool live_status = false;
+	unsigned int retry = 3;
 
 	DRM_DEBUG_KMS("[CONNECTOR:%d:%s]\n",
 		      connector->base.id, connector->name);
 
+	while (!live_status && --retry) {
+		live_status = intel_digital_port_connected(dev_priv,
+				hdmi_to_dig_port(intel_hdmi));
+		mdelay(10);
+	}
+
+	if (!live_status)
+		DRM_DEBUG_KMS("Live status not up!");
+
 	intel_hdmi_unset_edid(connector);
 
-	if (intel_hdmi_set_edid(connector)) {
+	if (intel_hdmi_set_edid(connector, live_status)) {
 		struct intel_hdmi *intel_hdmi = intel_attached_hdmi(connector);
 
 		hdmi_to_dig_port(intel_hdmi)->base.type = INTEL_OUTPUT_HDMI;
@@ -1402,7 +1416,7 @@ intel_hdmi_force(struct drm_connector *connector)
 	if (connector->status != connector_status_connected)
 		return;
 
-	intel_hdmi_set_edid(connector);
+	intel_hdmi_set_edid(connector, true);
 	hdmi_to_dig_port(intel_hdmi)->base.type = INTEL_OUTPUT_HDMI;
 }
 
Signature · 6 lines
1.7.10.4

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

Re: [PATCH] drm/i915: Check live status before reading edid

Quoted 62 lines · Sonika Jindal

The Bspec is very clear that Live status must be checked about before trying to read EDID over DDC channel. This patch makes sure that HDMI EDID is read only when live status is up.

The live status doesn't seem to perform very consistent across various platforms when tested with different monitors. The reason behind that is some monitors are late to provide right voltage to set live_status up. So, after getting the interrupt, for a small duration, live status reg fluctuates, and then settles down showing the correct staus.

This is explained here in, in a rough way:
HPD line ________________

			 |\ T1 = Monitor Hotplug causing IRQ
			 | \______________________________________
			 | |
                         | |
			 | |   T2 = Live status is stable
			 | |  _____________________________________
			 | | /|
Live status _____________|_|/ |
			 | |  |
			 | |  |
			 | |  |
			T0 T1  T2

(Between T1 and T2 Live status fluctuates or can be even low, depending on
the monitor)

After several experiments, we have concluded that a max delay of 30ms is enough to allow the live status to settle down with most of the monitors. This total delay of 30ms has been split into a resolution of 3 retries of 10ms each, for the better cases.

This delay is kept at 30ms, keeping in consideration that, HDCP compliance expect the HPD handler to respond a plug out in 100ms, by disabling port.

v2: Adding checks for VLV/CHV as well. Reusing old ibx and g4x functions to check digital port status. Adding a separate function to get bxt live status (Daniel)
v3: Using intel_encoder->hpd_pin to check the live status (Siva) Moving the live status read to intel_hdmi_probe and passing parameter to read/not to read the edid. (me)
v4:
* Added live status check for all platforms using intel_digital_port_connected.
* Rebased on top of Jani's DP cleanup series
* Some monitors take time in setting the live status. So retry for few times if this is a connect HPD
v5: Removed extra "drm/i915" from commit message. Adding Shashank's sob
which was missed.
v6: Drop the (!detect_edid && !live_status check) check because for DDI ports which are enumerated as hdmi as well as DP, we don't have a mechanism to differentiate between DP and hdmi inside the encoder's hot_plug. This leads to call to the hdmi's hot_plug hook for DP as well as hdmi which leads to issues during unplug because of the above check.
v7: Make intel_digital_port_connected global in this patch, some reformatting of while loop, adding a print when live status is not up. (Rodrigo)
v8: Rebase it on nightly which involved skipping the hot_plug hook for now and letting the live_status check happen in detect until the hpd handling part is finalized (Daniel)

Quoted 3 lines

Queued for -next, thanks for the patch.
-Daniel

Quoted 109 lines
---
 drivers/gpu/drm/i915/intel_dp.c   |    2 +-
 drivers/gpu/drm/i915/intel_drv.h  |    2 ++
 drivers/gpu/drm/i915/intel_hdmi.c |   28 +++++++++++++++++++++-------
 3 files changed, 24 insertions(+), 8 deletions(-)
drivers/gpu/drm/i915/intel_dp.c +1 -1
index a687250..a6e22dd 100644
@@ -4670,7 +4670,7 @@ static bool bxt_digital_port_connected(struct drm_i915_private *dev_priv,
  *
  * Return %true if @port is connected, %false otherwise.
  */
-static bool intel_digital_port_connected(struct drm_i915_private *dev_priv,
+bool intel_digital_port_connected(struct drm_i915_private *dev_priv,
 					 struct intel_digital_port *port)
 {
 	if (HAS_PCH_IBX(dev_priv))
drivers/gpu/drm/i915/intel_drv.h +2 -0
index 08ba362..1e01350 100644
@@ -1210,6 +1210,8 @@ void intel_edp_drrs_disable(struct intel_dp *intel_dp);
 void intel_edp_drrs_invalidate(struct drm_device *dev,
 		unsigned frontbuffer_bits);
 void intel_edp_drrs_flush(struct drm_device *dev, unsigned frontbuffer_bits);
+bool intel_digital_port_connected(struct drm_i915_private *dev_priv,
+					 struct intel_digital_port *port);
 void hsw_dp_set_ddi_pll_sel(struct intel_crtc_state *pipe_config);
 
 /* intel_dp_mst.c */
drivers/gpu/drm/i915/intel_hdmi.c +21 -7
index e978c59..bb33c66 100644
@@ -1329,22 +1329,23 @@ intel_hdmi_unset_edid(struct drm_connector *connector)
 }
 
 static bool
-intel_hdmi_set_edid(struct drm_connector *connector)
+intel_hdmi_set_edid(struct drm_connector *connector, bool force)
 {
 	struct drm_i915_private *dev_priv = to_i915(connector->dev);
 	struct intel_hdmi *intel_hdmi = intel_attached_hdmi(connector);
 	struct intel_encoder *intel_encoder =
 		&hdmi_to_dig_port(intel_hdmi)->base;
 	enum intel_display_power_domain power_domain;
-	struct edid *edid;
+	struct edid *edid = NULL;
 	bool connected = false;
 
 	power_domain = intel_display_port_power_domain(intel_encoder);
 	intel_display_power_get(dev_priv, power_domain);
 
-	edid = drm_get_edid(connector,
-			    intel_gmbus_get_adapter(dev_priv,
-						    intel_hdmi->ddc_bus));
+	if (force)
+		edid = drm_get_edid(connector,
+				    intel_gmbus_get_adapter(dev_priv,
+				    intel_hdmi->ddc_bus));
 
 	intel_display_power_put(dev_priv, power_domain);
 
@@ -1372,13 +1373,26 @@ static enum drm_connector_status
 intel_hdmi_detect(struct drm_connector *connector, bool force)
 {
 	enum drm_connector_status status;
+	struct intel_hdmi *intel_hdmi = intel_attached_hdmi(connector);
+	struct drm_i915_private *dev_priv = to_i915(connector->dev);
+	bool live_status = false;
+	unsigned int retry = 3;
 
 	DRM_DEBUG_KMS("[CONNECTOR:%d:%s]\n",
 		      connector->base.id, connector->name);
 
+	while (!live_status && --retry) {
+		live_status = intel_digital_port_connected(dev_priv,
+				hdmi_to_dig_port(intel_hdmi));
+		mdelay(10);
+	}
+
+	if (!live_status)
+		DRM_DEBUG_KMS("Live status not up!");
+
 	intel_hdmi_unset_edid(connector);
 
-	if (intel_hdmi_set_edid(connector)) {
+	if (intel_hdmi_set_edid(connector, live_status)) {
 		struct intel_hdmi *intel_hdmi = intel_attached_hdmi(connector);
 
 		hdmi_to_dig_port(intel_hdmi)->base.type = INTEL_OUTPUT_HDMI;
@@ -1402,7 +1416,7 @@ intel_hdmi_force(struct drm_connector *connector)
 	if (connector->status != connector_status_connected)
 		return;
 
-	intel_hdmi_set_edid(connector);
+	intel_hdmi_set_edid(connector, true);
 	hdmi_to_dig_port(intel_hdmi)->base.type = INTEL_OUTPUT_HDMI;
 }
 
Signature · 6 lines
1.7.10.4

_______________________________________________
Intel-gfx mailing list
[email protected]
http://lists.freedesktop.org/mailman/listinfo/intel-gfx
Signature · 7 lines
Daniel Vetter
Software Engineer, Intel Corporation
http://blog.ffwll.ch
_______________________________________________
Intel-gfx mailing list
[email protected]
http://lists.freedesktop.org/mailman/listinfo/intel-gfx
#52 lore raw

Re: [PATCH] drm/i915: Check live status before reading edid

In reply to Sonika Jindal
Quoted 28 lines · Sonika Jindal

The Bspec is very clear that Live status must be checked about before trying to read EDID over DDC channel. This patch makes sure that HDMI EDID is read only when live status is up.

The live status doesn't seem to perform very consistent across various platforms when tested with different monitors. The reason behind that is some monitors are late to provide right voltage to set live_status up. So, after getting the interrupt, for a small duration, live status reg fluctuates, and then settles down showing the correct staus.

This is explained here in, in a rough way:
HPD line ________________

			 |\ T1 = Monitor Hotplug causing IRQ
			 | \______________________________________
			 | |
                         | |
			 | |   T2 = Live status is stable
			 | |  _____________________________________
			 | | /|
Live status _____________|_|/ |
			 | |  |
			 | |  |
			 | |  |
			T0 T1  T2

(Between T1 and T2 Live status fluctuates or can be even low, depending on
the monitor)

Quoted 7 lines

After several experiments, we have concluded that a max delay of 30ms is enough to allow the live status to settle down with most of the monitors. This total delay of 30ms has been split into a resolution of 3 retries of 10ms each, for the better cases.

This delay is kept at 30ms, keeping in consideration that, HDCP compliance expect the HPD handler to respond a plug out in 100ms, by disabling port.

This is a regression-fest. Revert with stable@?
-Chris

Signature · 5 lines
Chris Wilson, Intel Open Source Technology Centre
_______________________________________________
Intel-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/intel-gfx
#53 lore raw

Re: [PATCH] drm/i915: Check live status before reading edid

+Shashank

Shashank was planning to give a patch to bypass live status checks for older platforms.

Regards,
Sonika

-----Original Message-----
From: Chris Wilson [mailto:[email protected]]
Sent: Wednesday, March 9, 2016 2:34 AM
To: Jindal, Sonika <[email protected]>
Cc: [email protected]
Subject: Re: [Intel-gfx] [PATCH] drm/i915: Check live status before reading edid

Quoted 28 lines · Sonika Jindal

The Bspec is very clear that Live status must be checked about before trying to read EDID over DDC channel. This patch makes sure that HDMI EDID is read only when live status is up.

The live status doesn't seem to perform very consistent across various platforms when tested with different monitors. The reason behind that is some monitors are late to provide right voltage to set live_status up. So, after getting the interrupt, for a small duration, live status reg fluctuates, and then settles down showing the correct staus.

This is explained here in, in a rough way:
HPD line ________________

			 |\ T1 = Monitor Hotplug causing IRQ
			 | \______________________________________
			 | |
                         | |
			 | |   T2 = Live status is stable
			 | |  _____________________________________
			 | | /|
Live status _____________|_|/ |
			 | |  |
			 | |  |
			 | |  |
			T0 T1  T2

(Between T1 and T2 Live status fluctuates or can be even low, depending on the monitor)

Quoted 7 lines

After several experiments, we have concluded that a max delay of 30ms is enough to allow the live status to settle down with most of the monitors. This total delay of 30ms has been split into a resolution of 3 retries of 10ms each, for the better cases.

This delay is kept at 30ms, keeping in consideration that, HDCP compliance expect the HPD handler to respond a plug out in 100ms, by disabling port.

This is a regression-fest. Revert with stable@?
-Chris

--
Chris Wilson, Intel Open Source Technology Centre
_______________________________________________
Intel-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/intel-gfx

#54 lore raw

Re: [PATCH] drm/i915: Check live status before reading edid

We are procuring the DVI-single-link cable, as we don't have one with us. Sonika tested the Dual link cable, and that was working well.

We can do two things here:
- Add the gen check, which will allow the live_status check only for VLV and + platforms, others will go as it is.
- Wait for some more time, get the cable and try to rep and fix the issue.

Regards
Shashank
-----Original Message-----
From: Jindal, Sonika
Sent: Wednesday, March 09, 2016 11:23 AM
To: Chris Wilson; Sharma, Shashank
Cc: [email protected]
Subject: RE: [Intel-gfx] [PATCH] drm/i915: Check live status before reading edid

+Shashank

Shashank was planning to give a patch to bypass live status checks for older platforms.

Regards,
Sonika

-----Original Message-----
From: Chris Wilson [mailto:[email protected]]
Sent: Wednesday, March 9, 2016 2:34 AM
To: Jindal, Sonika <[email protected]>
Cc: [email protected]
Subject: Re: [Intel-gfx] [PATCH] drm/i915: Check live status before reading edid

Quoted 28 lines · Sonika Jindal

The Bspec is very clear that Live status must be checked about before trying to read EDID over DDC channel. This patch makes sure that HDMI EDID is read only when live status is up.

The live status doesn't seem to perform very consistent across various platforms when tested with different monitors. The reason behind that is some monitors are late to provide right voltage to set live_status up. So, after getting the interrupt, for a small duration, live status reg fluctuates, and then settles down showing the correct staus.

This is explained here in, in a rough way:
HPD line ________________

			 |\ T1 = Monitor Hotplug causing IRQ
			 | \______________________________________
			 | |
                         | |
			 | |   T2 = Live status is stable
			 | |  _____________________________________
			 | | /|
Live status _____________|_|/ |
			 | |  |
			 | |  |
			 | |  |
			T0 T1  T2

(Between T1 and T2 Live status fluctuates or can be even low, depending on the monitor)

Quoted 7 lines

After several experiments, we have concluded that a max delay of 30ms is enough to allow the live status to settle down with most of the monitors. This total delay of 30ms has been split into a resolution of 3 retries of 10ms each, for the better cases.

This delay is kept at 30ms, keeping in consideration that, HDCP compliance expect the HPD handler to respond a plug out in 100ms, by disabling port.

This is a regression-fest. Revert with stable@?
-Chris

--
Chris Wilson, Intel Open Source Technology Centre
_______________________________________________
Intel-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/intel-gfx

#55 lore raw

Re: [PATCH] drm/i915: Check live status before reading edid

Hi all,

I have sent one patch with live_status check restricted to gen7 and +, for this regression. drm/i915: Restrict usage of live status check
Please help in review.

Regards
Shashank
-----Original Message-----
From: Sharma, Shashank
Sent: Wednesday, March 09, 2016 11:25 AM
To: Jindal, Sonika; Chris Wilson
Cc: [email protected]; Mukherjee, Indranil
Subject: RE: [Intel-gfx] [PATCH] drm/i915: Check live status before reading edid

We are procuring the DVI-single-link cable, as we don't have one with us. Sonika tested the Dual link cable, and that was working well.

We can do two things here:
- Add the gen check, which will allow the live_status check only for VLV and + platforms, others will go as it is.
- Wait for some more time, get the cable and try to rep and fix the issue.

Regards
Shashank
-----Original Message-----
From: Jindal, Sonika
Sent: Wednesday, March 09, 2016 11:23 AM
To: Chris Wilson; Sharma, Shashank
Cc: [email protected]
Subject: RE: [Intel-gfx] [PATCH] drm/i915: Check live status before reading edid

+Shashank

Shashank was planning to give a patch to bypass live status checks for older platforms.

Regards,
Sonika

-----Original Message-----
From: Chris Wilson [mailto:[email protected]]
Sent: Wednesday, March 9, 2016 2:34 AM
To: Jindal, Sonika <[email protected]>
Cc: [email protected]
Subject: Re: [Intel-gfx] [PATCH] drm/i915: Check live status before reading edid

Quoted 28 lines · Sonika Jindal

The Bspec is very clear that Live status must be checked about before trying to read EDID over DDC channel. This patch makes sure that HDMI EDID is read only when live status is up.

The live status doesn't seem to perform very consistent across various platforms when tested with different monitors. The reason behind that is some monitors are late to provide right voltage to set live_status up. So, after getting the interrupt, for a small duration, live status reg fluctuates, and then settles down showing the correct staus.

This is explained here in, in a rough way:
HPD line ________________

			 |\ T1 = Monitor Hotplug causing IRQ
			 | \______________________________________
			 | |
                         | |
			 | |   T2 = Live status is stable
			 | |  _____________________________________
			 | | /|
Live status _____________|_|/ |
			 | |  |
			 | |  |
			 | |  |
			T0 T1  T2

(Between T1 and T2 Live status fluctuates or can be even low, depending on the monitor)

Quoted 7 lines

After several experiments, we have concluded that a max delay of 30ms is enough to allow the live status to settle down with most of the monitors. This total delay of 30ms has been split into a resolution of 3 retries of 10ms each, for the better cases.

This delay is kept at 30ms, keeping in consideration that, HDCP compliance expect the HPD handler to respond a plug out in 100ms, by disabling port.

This is a regression-fest. Revert with stable@?
-Chris

--
Chris Wilson, Intel Open Source Technology Centre
_______________________________________________
Intel-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/intel-gfx

#56 lore raw

PATCH 5/6 drm/i915: drm/i915: Process hpd only for hdmi inside hotplug_work_func

In reply to Sonika Jindal

If the same port is enumerated as hdmi as well as DP, this will get called for DP connector as well which is not required because i915_hotplug_work_func is solely to handle hdmi HPD.

---
 drivers/gpu/drm/i915/intel_hotplug.c |    3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)
drivers/gpu/drm/i915/intel_hotplug.c +2 -1
index 53c0173..8e1c43e 100644
@@ -325,7 +325,8 @@ static void i915_hotplug_work_func(struct work_struct *work)
 
 	list_for_each_entry(connector, &mode_config->connector_list, head) {
 		intel_connector = to_intel_connector(connector);
-		if (!intel_connector->encoder)
+		if (!intel_connector->encoder
+			&& connector->connector_type != DRM_MODE_CONNECTOR_HDMIA)
 			continue;
 		intel_encoder = intel_connector->encoder;
 		if (hpd_event_bits & (1 << intel_encoder->hpd_pin)) {
Signature · 6 lines
1.7.10.4

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

Re: [PATCH 5/6] drm/i915: drm/i915: Process hpd only for hdmi inside hotplug_work_func

Quoted 14 lines · Sonika Jindal

If the same port is enumerated as hdmi as well as DP, this will get called for DP connector as well which is not required because i915_hotplug_work_func is solely to handle hdmi HPD.

---
 drivers/gpu/drm/i915/intel_hotplug.c |    3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)
drivers/gpu/drm/i915/intel_hotplug.c +0 -0
index 53c0173..8e1c43e 100644

drivers/gpu/drm/i915/intel_hotplug.c static void i915_hotplug_work_func(struct work_struct *work) · line 330

Quoted 7 lines
drivers/gpu/drm/i915/intel_hotplug.c +2 -1
@@ -325,7 +325,8 @@ static void i915_hotplug_work_func(struct work_struct *work)
 
 	list_for_each_entry(connector, &mode_config->connector_list, head) {
 		intel_connector = to_intel_connector(connector);
-		if (!intel_connector->encoder)
+		if (!intel_connector->encoder
+			&& connector->connector_type != DRM_MODE_CONNECTOR_HDMIA)
 			continue;

Pretty sure this breaks hotplug detection for everything but HDMI (since now nothing but hdmi gets called it's ->detect function, and only if that signals a status change will we send out an uevent to userspace). Does this really not break DP hotplug?

Yes we probe both hdmi and DP always, and probably we could be more intelligent with the fallback between the too. But this here doesn't seem to be the right solution.
-Daniel

Quoted 9 lines
drivers/gpu/drm/i915/intel_hotplug.c +0 -0
 		intel_encoder = intel_connector->encoder;
 		if (hpd_event_bits & (1 << intel_encoder->hpd_pin)) {
Signature · 6 lines
1.7.10.4

_______________________________________________
Intel-gfx mailing list
[email protected]
http://lists.freedesktop.org/mailman/listinfo/intel-gfx
Signature · 7 lines
Daniel Vetter
Software Engineer, Intel Corporation
http://blog.ffwll.ch
_______________________________________________
Intel-gfx mailing list
[email protected]
http://lists.freedesktop.org/mailman/listinfo/intel-gfx
#58 lore raw

Re: [PATCH 5/6] drm/i915: drm/i915: Process hpd only for hdmi inside hotplug_work_func

Quoted 28 lines · Daniel Vetter
Quoted 14 lines · Sonika Jindal

If the same port is enumerated as hdmi as well as DP, this will get called for DP connector as well which is not required because i915_hotplug_work_func is solely to handle hdmi HPD.

---
  drivers/gpu/drm/i915/intel_hotplug.c |    3 ++-
  1 file changed, 2 insertions(+), 1 deletion(-)
drivers/gpu/drm/i915/intel_hotplug.c +0 -0
index 53c0173..8e1c43e 100644

drivers/gpu/drm/i915/intel_hotplug.c static void i915_hotplug_work_func(struct work_struct *work) · line 330

Quoted 7 lines
drivers/gpu/drm/i915/intel_hotplug.c +2 -1
@@ -325,7 +325,8 @@ static void i915_hotplug_work_func(struct work_struct *work)
 
  	list_for_each_entry(connector, &mode_config->connector_list, head) {
  		intel_connector = to_intel_connector(connector);
-		if (!intel_connector->encoder)
+		if (!intel_connector->encoder
+			&& connector->connector_type != DRM_MODE_CONNECTOR_HDMIA)
  			continue;

Pretty sure this breaks hotplug detection for everything but HDMI (since now nothing but hdmi gets called it's ->detect function, and only if that signals a status change will we send out an uevent to userspace). Does this really not break DP hotplug?

Quoted 4 lines

Yes we probe both hdmi and DP always, and probably we could be more intelligent with the fallback between the too. But this here doesn't seem to be the right solution.
-Daniel

Hmm :(
This doesn't allow detect to be called for dp. I missed the part that hpd_pulse only handles mst. From the name and comments it looks like it should handle hpd for dp completely. I think this this code needs some refactoring.

For now, I think I better move this check to intel_encoder->hot_plug function because for the connectors with dp and hdmi both, this hot_plug hook gets called twice.

Quoted 11 lines
Quoted 10 lines
drivers/gpu/drm/i915/intel_hotplug.c +0 -1
  		intel_encoder = intel_connector->encoder;
  		if (hpd_event_bits & (1 << intel_encoder->hpd_pin)) {
--

1.7.10.4

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

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

#59 lore raw

PATCH 1/1 drm/i915: Call encoder hot_plug hook only for hdmi

If the same port is enumerated as hdmi as well as DP, this will call hot_plug hook for DP as well which is not required here. This splitting of edid read and detect is done only for HDMI with this series.

v2: Avoid breaking DP hpd. Also corrected the commit message and description accordingly. (Daniel)

---
 drivers/gpu/drm/i915/intel_hotplug.c |    3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)
drivers/gpu/drm/i915/intel_hotplug.c +2 -1
index 53c0173..ff4692a 100644
@@ -331,7 +331,8 @@ static void i915_hotplug_work_func(struct work_struct *work)
 		if (hpd_event_bits & (1 << intel_encoder->hpd_pin)) {
 			DRM_DEBUG_KMS("Connector %s (pin %i) received hotplug event.\n",
 				      connector->name, intel_encoder->hpd_pin);
-			if (intel_encoder->hot_plug)
+			if (intel_encoder->hot_plug
+				&& connector->connector_type == DRM_MODE_CONNECTOR_HDMIA)
 				intel_encoder->hot_plug(intel_encoder);
 			if (intel_hpd_irq_event(dev, connector))
 				changed = true;
Signature · 6 lines
1.7.10.4

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

Re: [PATCH] drm/i915: Call encoder hot_plug hook only for hdmi

Quoted 18 lines · Sonika Jindal

If the same port is enumerated as hdmi as well as DP, this will call hot_plug hook for DP as well which is not required here. This splitting of edid read and detect is done only for HDMI with this series.

v2: Avoid breaking DP hpd. Also corrected the commit message and description accordingly. (Daniel)

---
 drivers/gpu/drm/i915/intel_hotplug.c |    3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)
drivers/gpu/drm/i915/intel_hotplug.c +0 -0
index 53c0173..ff4692a 100644

drivers/gpu/drm/i915/intel_hotplug.c static void i915_hotplug_work_func(struct work_struct *work) · line 335

Quoted 6 lines
drivers/gpu/drm/i915/intel_hotplug.c +2 -1
@@ -331,7 +331,8 @@ static void i915_hotplug_work_func(struct work_struct *work)
 		if (hpd_event_bits & (1 << intel_encoder->hpd_pin)) {
 			DRM_DEBUG_KMS("Connector %s (pin %i) received hotplug event.\n",
 				      connector->name, intel_encoder->hpd_pin);
-			if (intel_encoder->hot_plug)
+			if (intel_encoder->hot_plug
+				&& connector->connector_type == DRM_MODE_CONNECTOR_HDMIA)

Please use something like grep to find all the other ->hot_plug implementations and then please tell me why you don't break them all. -Daniel

Quoted 10 lines
drivers/gpu/drm/i915/intel_hotplug.c +0 -0
 				intel_encoder->hot_plug(intel_encoder);
 			if (intel_hpd_irq_event(dev, connector))
 				changed = true;
Signature · 6 lines
1.7.10.4

_______________________________________________
Intel-gfx mailing list
[email protected]
http://lists.freedesktop.org/mailman/listinfo/intel-gfx
Signature · 7 lines
Daniel Vetter
Software Engineer, Intel Corporation
http://blog.ffwll.ch
_______________________________________________
Intel-gfx mailing list
[email protected]
http://lists.freedesktop.org/mailman/listinfo/intel-gfx
#61 lore raw

Re: [PATCH] drm/i915: Call encoder hot_plug hook only for hdmi

Quoted 14 lines · Daniel Vetter
Quoted 13 lines · Sonika Jindal

If the same port is enumerated as hdmi as well as DP, this will call hot_plug hook for DP as well which is not required here. This splitting of edid read and detect is done only for HDMI with this series.

v2: Avoid breaking DP hpd. Also corrected the commit message and description accordingly. (Daniel)

---
  drivers/gpu/drm/i915/intel_hotplug.c |    3 ++-
  1 file changed, 2 insertions(+), 1 deletion(-)
Quoted 16 lines

drivers/gpu/drm/i915/intel_hotplug.c static void i915_hotplug_work_func(struct work_struct *work) · line 335

Quoted 11 lines
drivers/gpu/drm/i915/intel_hotplug.c +2 -1
index 53c0173..ff4692a 100644
@@ -331,7 +331,8 @@ static void i915_hotplug_work_func(struct work_struct *work)
  		if (hpd_event_bits & (1 << intel_encoder->hpd_pin)) {
  			DRM_DEBUG_KMS("Connector %s (pin %i) received hotplug event.\n",
  				      connector->name, intel_encoder->hpd_pin);
-			if (intel_encoder->hot_plug)
+			if (intel_encoder->hot_plug
+				&& connector->connector_type == DRM_MODE_CONNECTOR_HDMIA)

Please use something like grep to find all the other ->hot_plug implementations and then please tell me why you don't break them all. -Daniel

Hmm, I only checked for hot_plug for DP/edp which is not there. Failed to notice that there is one in intel_sdvo.c. My mistake. I will place it properly somewhere else.

Regards,
Sonika

Quoted 11 lines
Quoted 10 lines
drivers/gpu/drm/i915/intel_hotplug.c +0 -1
  				intel_encoder->hot_plug(intel_encoder);
  			if (intel_hpd_irq_event(dev, connector))
  				changed = true;
--

1.7.10.4

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

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

#62 lore raw

Re: [PATCH] drm/i915: Call encoder hot_plug hook only for hdmi

Quoted 37 lines · Sonika
Quoted 25 lines · Daniel Vetter
Quoted 24 lines · Sonika Jindal

If the same port is enumerated as hdmi as well as DP, this will call hot_plug hook for DP as well which is not required here. This splitting of edid read and detect is done only for HDMI with this series.

v2: Avoid breaking DP hpd. Also corrected the commit message and description accordingly. (Daniel)

---
  drivers/gpu/drm/i915/intel_hotplug.c |    3 ++-
  1 file changed, 2 insertions(+), 1 deletion(-)
drivers/gpu/drm/i915/intel_hotplug.c +0 -0
index 53c0173..ff4692a 100644
@@ -331,7 +331,8 @@ static void i915_hotplug_work_func(struct
 work_struct *work)
          if (hpd_event_bits & (1 << intel_encoder->hpd_pin)) {
              DRM_DEBUG_KMS("Connector %s (pin %i) received hotplug
 event.\n",
                        connector->name, intel_encoder->hpd_pin);
Quoted 9 lines

drivers/gpu/drm/i915/intel_hotplug.c static void i915_hotplug_work_func(struct work_struct *work)

Quoted 4 lines
drivers/gpu/drm/i915/intel_hotplug.c +2 -1
-            if (intel_encoder->hot_plug)
+            if (intel_encoder->hot_plug
+                && connector->connector_type ==
 DRM_MODE_CONNECTOR_HDMIA)

Please use something like grep to find all the other ->hot_plug implementations and then please tell me why you don't break them all. -Daniel

Quoted 6 lines

Hmm, I only checked for hot_plug for DP/edp which is not there. Failed to notice that there is one in intel_sdvo.c.
My mistake. I will place it properly somewhere else.

Regards,
Sonika

Is there any suggestion about how we can differentiate if it is actual DP or HDMI hotplug at this point? intel_encoder's type gets updated after detect call. So, not sure how to have this kind of check.

For now, I think we can abandon this patch from this series.

Regards,
Sonika

Quoted 11 lines
Quoted 11 lines
Quoted 10 lines
                  intel_encoder->hot_plug(intel_encoder);
              if (intel_hpd_irq_event(dev, connector))
                  changed = true;

--
1.7.10.4

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

Quoted 4 lines

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

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

#63 lore raw

Re: [PATCH] drm/i915: Call encoder hot_plug hook only for hdmi

Quoted 44 lines · Sonika
Quoted 41 lines · Sonika
Quoted 25 lines · Daniel Vetter
Quoted 24 lines · Sonika Jindal

If the same port is enumerated as hdmi as well as DP, this will call hot_plug hook for DP as well which is not required here. This splitting of edid read and detect is done only for HDMI with this series.

v2: Avoid breaking DP hpd. Also corrected the commit message and description accordingly. (Daniel)

---
 drivers/gpu/drm/i915/intel_hotplug.c |    3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)
drivers/gpu/drm/i915/intel_hotplug.c +0 -0
index 53c0173..ff4692a 100644
@@ -331,7 +331,8 @@ static void i915_hotplug_work_func(struct
 work_struct *work)
         if (hpd_event_bits & (1 << intel_encoder->hpd_pin)) {
             DRM_DEBUG_KMS("Connector %s (pin %i) received hotplug
 event.\n",
                       connector->name, intel_encoder->hpd_pin);
Quoted 9 lines

drivers/gpu/drm/i915/intel_hotplug.c static void i915_hotplug_work_func(struct work_struct *work)

Quoted 4 lines
drivers/gpu/drm/i915/intel_hotplug.c +2 -1
-            if (intel_encoder->hot_plug)
+            if (intel_encoder->hot_plug
+                && connector->connector_type ==
 DRM_MODE_CONNECTOR_HDMIA)

Please use something like grep to find all the other ->hot_plug implementations and then please tell me why you don't break them all. -Daniel

Hmm, I only checked for hot_plug for DP/edp which is not there. Failed to notice that there is one in intel_sdvo.c.
My mistake. I will place it properly somewhere else.

Quoted 8 lines
Quoted 2 lines

Regards,
Sonika

Is there any suggestion about how we can differentiate if it is actual DP or HDMI hotplug at this point? intel_encoder's type gets updated after detect call. So, not sure how to have this kind of check.

For now, I think we can abandon this patch from this series.

No, hpd is shared between hdmi and dp at the hw level so we can't differentiate. Long term my idea would be that we merge together all the hdmi _and_ dp hpd handling into one overall control-flow. Then we can make sure to not call anything twice and also have sensible high-level flow (like first checking for dp vs. hdmi and then taking relevant paths).

For dealing with ->hot_plug a quick fix might be to have a separate loop going over encoders (to make sure we only call it once per encoder if there's more than one connector for 1 encoder). That behaviour would also be ok for sdvo.
-Daniel

Signature · 7 lines
Daniel Vetter
Software Engineer, Intel Corporation
http://blog.ffwll.ch
_______________________________________________
Intel-gfx mailing list
[email protected]
http://lists.freedesktop.org/mailman/listinfo/intel-gfx
#64 lore raw

Re: [PATCH] drm/i915: Call encoder hot_plug hook only for hdmi

-----Original Message-----
From: Daniel Vetter [mailto:[email protected]] On Behalf Of Daniel Vetter
Sent: Wednesday, September 9, 2015 8:48 PM
To: Jindal, Sonika
Cc: Daniel Vetter; [email protected]; Sharma, Shashank
Subject: Re: [Intel-gfx] [PATCH] drm/i915: Call encoder hot_plug hook only for hdmi

Quoted 44 lines · Sonika
Quoted 41 lines · Sonika
Quoted 25 lines · Daniel Vetter
Quoted 24 lines · Sonika Jindal

If the same port is enumerated as hdmi as well as DP, this will call hot_plug hook for DP as well which is not required here. This splitting of edid read and detect is done only for HDMI with this series.

v2: Avoid breaking DP hpd. Also corrected the commit message and description accordingly. (Daniel)

---
 drivers/gpu/drm/i915/intel_hotplug.c |    3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)
drivers/gpu/drm/i915/intel_hotplug.c +0 -0
index 53c0173..ff4692a 100644
@@ -331,7 +331,8 @@ static void i915_hotplug_work_func(struct 
 work_struct *work)
         if (hpd_event_bits & (1 << intel_encoder->hpd_pin)) {
             DRM_DEBUG_KMS("Connector %s (pin %i) received hotplug 
 event.\n",
                       connector->name, intel_encoder->hpd_pin);
Quoted 9 lines

drivers/gpu/drm/i915/intel_hotplug.c static void i915_hotplug_work_func(struct work_struct *work)

Quoted 4 lines
drivers/gpu/drm/i915/intel_hotplug.c +2 -1
-            if (intel_encoder->hot_plug)
+            if (intel_encoder->hot_plug
+                && connector->connector_type ==
 DRM_MODE_CONNECTOR_HDMIA)

Please use something like grep to find all the other ->hot_plug implementations and then please tell me why you don't break them all. -Daniel

Hmm, I only checked for hot_plug for DP/edp which is not there. Failed to notice that there is one in intel_sdvo.c. My mistake. I will place it properly somewhere else.

Quoted 8 lines
Quoted 2 lines

Regards,
Sonika

Is there any suggestion about how we can differentiate if it is actual DP or HDMI hotplug at this point? intel_encoder's type gets updated after detect call. So, not sure how to have this kind of check.

For now, I think we can abandon this patch from this series.

No, hpd is shared between hdmi and dp at the hw level so we can't differentiate. Long term my idea would be that we merge together all the hdmi _and_ dp hpd handling into one overall control-flow. Then we can make sure to not call anything twice and also have sensible high-level flow (like first checking for dp vs. hdmi and then taking relevant paths).

For dealing with ->hot_plug a quick fix might be to have a separate loop going over encoders (to make sure we only call it once per encoder if there's more than one connector for 1 encoder). That behaviour would also be ok for sdvo.

<Sonika> Hmm, so instead of relying on connector, we can check for the hpd_pin on encoder and remove the connector loop completely?

-Daniel
--
Daniel Vetter
Software Engineer, Intel Corporation
http://blog.ffwll.ch
_______________________________________________
Intel-gfx mailing list
[email protected]
http://lists.freedesktop.org/mailman/listinfo/intel-gfx

#65 lore raw

Re: [PATCH] drm/i915: Call encoder hot_plug hook only for hdmi

Quoted 70 lines · Sonika

-----Original Message-----
From: Daniel Vetter [mailto:[email protected]] On Behalf Of Daniel Vetter
Sent: Wednesday, September 9, 2015 8:48 PM
To: Jindal, Sonika
Cc: Daniel Vetter; [email protected]; Sharma, Shashank
Subject: Re: [Intel-gfx] [PATCH] drm/i915: Call encoder hot_plug hook only for hdmi

Quoted 44 lines · Sonika
Quoted 41 lines · Sonika
Quoted 25 lines · Daniel Vetter
Quoted 24 lines · Sonika Jindal

If the same port is enumerated as hdmi as well as DP, this will call hot_plug hook for DP as well which is not required here. This splitting of edid read and detect is done only for HDMI with this series.

v2: Avoid breaking DP hpd. Also corrected the commit message and description accordingly. (Daniel)

---
 drivers/gpu/drm/i915/intel_hotplug.c |    3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)
drivers/gpu/drm/i915/intel_hotplug.c +0 -0
index 53c0173..ff4692a 100644
@@ -331,7 +331,8 @@ static void i915_hotplug_work_func(struct 
 work_struct *work)
         if (hpd_event_bits & (1 << intel_encoder->hpd_pin)) {
             DRM_DEBUG_KMS("Connector %s (pin %i) received hotplug 
 event.\n",
                       connector->name, intel_encoder->hpd_pin);
Quoted 9 lines

drivers/gpu/drm/i915/intel_hotplug.c static void i915_hotplug_work_func(struct work_struct *work)

Quoted 4 lines
drivers/gpu/drm/i915/intel_hotplug.c +2 -1
-            if (intel_encoder->hot_plug)
+            if (intel_encoder->hot_plug
+                && connector->connector_type ==
 DRM_MODE_CONNECTOR_HDMIA)

Please use something like grep to find all the other ->hot_plug implementations and then please tell me why you don't break them all. -Daniel

Hmm, I only checked for hot_plug for DP/edp which is not there. Failed to notice that there is one in intel_sdvo.c. My mistake. I will place it properly somewhere else.

Quoted 8 lines
Quoted 2 lines

Regards,
Sonika

Is there any suggestion about how we can differentiate if it is actual DP or HDMI hotplug at this point? intel_encoder's type gets updated after detect call. So, not sure how to have this kind of check.

For now, I think we can abandon this patch from this series.

No, hpd is shared between hdmi and dp at the hw level so we can't differentiate. Long term my idea would be that we merge together all the hdmi _and_ dp hpd handling into one overall control-flow. Then we can make sure to not call anything twice and also have sensible high-level flow (like first checking for dp vs. hdmi and then taking relevant paths).

Quoted 7 lines

For dealing with ->hot_plug a quick fix might be to have a separate loop going over encoders (to make sure we only call it once per encoder if there's more than one connector for 1 encoder). That behaviour would also be ok for sdvo.

<Sonika> Hmm, so instead of relying on connector, we can check for the hpd_pin on encoder and remove the connector loop completely?

That would be my suggestion, but I didn't check the details. -Daniel

Signature · 7 lines
Daniel Vetter
Software Engineer, Intel Corporation
http://blog.ffwll.ch
_______________________________________________
Intel-gfx mailing list
[email protected]
http://lists.freedesktop.org/mailman/listinfo/intel-gfx
#66 lore raw

Re: [PATCH 5/6] drm/i915: drm/i915: Process hpd only for hdmi inside hotplug_work_func

In reply to Jindal, Sonika

I was also going to say that I believe this breaks DP hotplug, but this was agreed already...

But I also don't understand how hot_plug is called twice since it calls encoder->hot_plug...
and dp has no ->hot_plug...
And also if this is happening how this code that checks for connector would help... there is something else strange happening...

Quoted 61 lines · Sonika
Quoted 11 lines · Daniel Vetter

Review

Quoted 10 lines · Sonika Jindal

If the same port is enumerated as hdmi as well as DP, this will get called for DP connector as well which is not required because i915_hotplug_work_func is solely to handle hdmi HPD.

---
  drivers/gpu/drm/i915/intel_hotplug.c |    3 ++-
  1 file changed, 2 insertions(+), 1 deletion(-)
(quoted diff) +0 -0

b/drivers/gpu/drm/i915/intel_hotplug.c

Quoted 4 lines

drivers/gpu/drm/i915/intel_hotplug.c static void i915_hotplug_work_func(struct

Quoted 4 lines

index 53c0173..8e1c43e 100644

drivers/gpu/drm/i915/intel_hotplug.c +0 -0
@@ -325,7 +325,8 @@ static void i915_hotplug_work_func(struct

work_struct *work)

Quoted 2 lines

drivers/gpu/drm/i915/intel_hotplug.c static void i915_hotplug_work_func(struct · line 325

Quoted 2 lines
drivers/gpu/drm/i915/intel_hotplug.c +0 -0
     list_for_each_entry(connector, &mode_config->connector_list, head)

{

Quoted 4 lines

drivers/gpu/drm/i915/intel_hotplug.c static void i915_hotplug_work_func(struct work_struct *work)

Quoted 4 lines
drivers/gpu/drm/i915/intel_hotplug.c +2 -1
             intel_connector = to_intel_connector(connector);
-            if (!intel_connector->encoder)
+            if (!intel_connector->encoder
+                    && connector->connector_type !=

DRM_MODE_CONNECTOR_HDMIA)

Quoted 11 lines

drivers/gpu/drm/i915/intel_hotplug.c static void i915_hotplug_work_func(struct work_struct *work) · line 330

Quoted 1 line
drivers/gpu/drm/i915/intel_hotplug.c +0 -0
                     continue;

Pretty sure this breaks hotplug detection for everything but HDMI (since now nothing but hdmi gets called it's ->detect function, and only if that signals a status change will we send out an uevent to userspace). Does this really not break DP hotplug?

Yes we probe both hdmi and DP always, and probably we could be more intelligent with the fallback between the too. But this here doesn't seem to be the right solution.
-Daniel

Hmm :(
This doesn't allow detect to be called for dp. I missed the part that hpd_pulse only handles mst. From the name and comments it looks like it should handle hpd for dp completely. I think this this code needs some refactoring.

For now, I think I better move this check to intel_encoder->hot_plug function because for the connectors with dp and hdmi both, this hot_plug hook gets called twice.

Quoted 11 lines
Quoted 10 lines
drivers/gpu/drm/i915/intel_hotplug.c +0 -1
             intel_encoder = intel_connector->encoder;
             if (hpd_event_bits & (1 << intel_encoder->hpd_pin)) {
--

1.7.10.4

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

Quoted 5 lines

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

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

#67 lore raw

Re: [PATCH 5/6] drm/i915: drm/i915: Process hpd only for hdmi inside hotplug_work_func

Since, the port is enumerated as DP as well as hdmi, the init connector is called for both and the same intel encoder gets attached to both the connectors.
Now the intel_encoder’s type could be a differentiator. But since, hdmi’s detect gets called, the type of the encoder remains hdmi.
When hpd occurs, it checks from the list of connectors whose hpd pin is same as the one asserted.
Since both dp and hdmi are INITed, both the connectors have that pin and since DP’s connector is first in the list, it calls the ->hot_plug once when it finds DP connector matching and then again when HDMI connector is found matching as per the pin.

This code check will avoid calling in the case when we have found the DP connector because we are sure that there is no ->hot_plug hook for DP. Anyways, that was not a generic implementation.

Regards,
Sonika

From: Rodrigo Vivi [mailto:[email protected]]
Sent: Thursday, September 10, 2015 12:51 AM
To: Jindal, Sonika; Daniel Vetter
Cc: [email protected]
Subject: Re: [Intel-gfx] [PATCH 5/6] drm/i915: drm/i915: Process hpd only for hdmi inside hotplug_work_func

I was also going to say that I believe this breaks DP hotplug, but this was agreed already...

But I also don't understand how hot_plug is called twice since it calls encoder->hot_plug...
and dp has no ->hot_plug...
And also if this is happening how this code that checks for connector would help... there is something else strange happening...

On Sat, Sep 5, 2015 at 9:31 PM Jindal, Sonika <[email protected]<mailto:[email protected]>> wrote:

Quoted 28 lines · Daniel Vetter
Quoted 14 lines · Sonika Jindal

If the same port is enumerated as hdmi as well as DP, this will get called for DP connector as well which is not required because i915_hotplug_work_func is solely to handle hdmi HPD.

---
  drivers/gpu/drm/i915/intel_hotplug.c |    3 ++-
  1 file changed, 2 insertions(+), 1 deletion(-)
drivers/gpu/drm/i915/intel_hotplug.c +0 -0
index 53c0173..8e1c43e 100644

drivers/gpu/drm/i915/intel_hotplug.c static void i915_hotplug_work_func(struct work_struct *work) · line 330

Quoted 7 lines
drivers/gpu/drm/i915/intel_hotplug.c +2 -1
@@ -325,7 +325,8 @@ static void i915_hotplug_work_func(struct work_struct *work)
 
     list_for_each_entry(connector, &mode_config->connector_list, head) {
             intel_connector = to_intel_connector(connector);
-            if (!intel_connector->encoder)
+            if (!intel_connector->encoder
+                    && connector->connector_type != DRM_MODE_CONNECTOR_HDMIA)
                     continue;

Pretty sure this breaks hotplug detection for everything but HDMI (since now nothing but hdmi gets called it's ->detect function, and only if that signals a status change will we send out an uevent to userspace). Does this really not break DP hotplug?

Quoted 4 lines

Yes we probe both hdmi and DP always, and probably we could be more intelligent with the fallback between the too. But this here doesn't seem to be the right solution.
-Daniel

Hmm :(
This doesn't allow detect to be called for dp. I missed the part that hpd_pulse only handles mst. From the name and comments it looks like it should handle hpd for dp completely. I think this this code needs some refactoring.

For now, I think I better move this check to intel_encoder->hot_plug function because for the connectors with dp and hdmi both, this hot_plug hook gets called twice.

Quoted 11 lines
Quoted 10 lines
drivers/gpu/drm/i915/intel_hotplug.c +0 -1
             intel_encoder = intel_connector->encoder;
             if (hpd_event_bits & (1 << intel_encoder->hpd_pin)) {
--

1.7.10.4

_______________________________________________
Intel-gfx mailing list [email protected]<mailto:[email protected]>
http://lists.freedesktop.org/mailman/listinfo/intel-gfx

_______________________________________________
Intel-gfx mailing list [email protected]<mailto:[email protected]>
http://lists.freedesktop.org/mailman/listinfo/intel-gfx

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

#68 lore raw

PATCH 6/6 drm/i915/bxt: Fix irq_port for eDP

In reply to Sonika Jindal

From: Durgadoss R <[email protected]>

Currently, HDMI hotplug with eDP as local panel is failing because the HDMI hpd is detected as a long hpd for eDP; and is thus rightfully ignored. But, it should really be handled as an interrupt on port B for HDMI (due to BXT A1 platform having HPD pins A and B swapped). This patch sets the irq_port[PORT_A] to NULL in case eDP is on port A so that irq handler does not treat it as a 'dig_port' interrupt.

---
 drivers/gpu/drm/i915/intel_ddi.c |   15 ++++++++++-----
 1 file changed, 10 insertions(+), 5 deletions(-)
drivers/gpu/drm/i915/intel_ddi.c +10 -5
index 4823184..fec51df 100644
@@ -3218,15 +3218,20 @@ void intel_ddi_init(struct drm_device *dev, enum port port)
 			goto err;
 
 		intel_dig_port->hpd_pulse = intel_dp_hpd_pulse;
+		dev_priv->hotplug.irq_port[port] = intel_dig_port;
 		/*
 		 * On BXT A0/A1, sw needs to activate DDIA HPD logic and
 		 * interrupts to check the external panel connection.
+		 * If eDP is connected on port A, set irq_port to NULL
+		 * so that we do not assume an interrupt here as a
+		 * 'dig_port' interrupt.
 		 */
-		if (IS_BROXTON(dev_priv) && (INTEL_REVID(dev) < BXT_REVID_B0)
-					 && port == PORT_B)
-			dev_priv->hotplug.irq_port[PORT_A] = intel_dig_port;
-		else
-			dev_priv->hotplug.irq_port[port] = intel_dig_port;
+		if (IS_BROXTON(dev) && (INTEL_REVID(dev) < BXT_REVID_B0)) {
+			if (port == PORT_B)
+				dev_priv->hotplug.irq_port[PORT_A] = intel_dig_port;
+			else if (intel_encoder->type == INTEL_OUTPUT_EDP)
+				dev_priv->hotplug.irq_port[port] = NULL;
+		}
 	}
 
 	/* In theory we don't need the encoder->type check, but leave it just in
Signature · 6 lines
1.7.10.4

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

Re: [PATCH 6/6] drm/i915/bxt: Fix irq_port for eDP

Nak: I don't believe we need this... Actually I believe we need the opposite... we need to enable HPD on port A for eDP errors handling...

Quoted 54 lines · Sonika Jindal

From: Durgadoss R <[email protected]>

Currently, HDMI hotplug with eDP as local panel is failing because the HDMI hpd is detected as a long hpd for eDP; and is thus rightfully ignored. But, it should really be handled as an interrupt on port B for HDMI (due to BXT A1 platform having HPD pins A and B swapped). This patch sets the irq_port[PORT_A] to NULL in case eDP is on port A so that irq handler does not treat it as a 'dig_port' interrupt.

---
 drivers/gpu/drm/i915/intel_ddi.c |   15 ++++++++++-----
 1 file changed, 10 insertions(+), 5 deletions(-)
drivers/gpu/drm/i915/intel_ddi.c +7 -5
index 4823184..fec51df 100644
@@ -3218,15 +3218,20 @@ void intel_ddi_init(struct drm_device *dev, enum
 port port)
                        goto err;
 
                intel_dig_port->hpd_pulse = intel_dp_hpd_pulse;
+               dev_priv->hotplug.irq_port[port] = intel_dig_port;
                /*
                 * On BXT A0/A1, sw needs to activate DDIA HPD logic and
                 * interrupts to check the external panel connection.
+                * If eDP is connected on port A, set irq_port to NULL
+                * so that we do not assume an interrupt here as a
+                * 'dig_port' interrupt.
                 */
-               if (IS_BROXTON(dev_priv) && (INTEL_REVID(dev) <
 BXT_REVID_B0)
-                                        && port == PORT_B)
-                       dev_priv->hotplug.irq_port[PORT_A] =
 intel_dig_port;
-               else
-                       dev_priv->hotplug.irq_port[port] = intel_dig_port;
+               if (IS_BROXTON(dev) && (INTEL_REVID(dev) < BXT_REVID_B0)) {
+                       if (port == PORT_B)
+                               dev_priv->hotplug.irq_port[PORT_A] =

intel_dig_port;

drivers/gpu/drm/i915/intel_ddi.c +3 -0
+                       else if (intel_encoder->type == INTEL_OUTPUT_EDP)
+                               dev_priv->hotplug.irq_port[port] = NULL;
+               }
        }
 
        /* In theory we don't need the encoder->type check, but leave it

just in
--
1.7.10.4

Quoted 5 lines

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

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

author NAK
#70 lore raw

Re: [PATCH 6/6] drm/i915/bxt: Fix irq_port for eDP

All this is because of the BXT A0/A1 workaround for HPD pins.
As you can see, this is only done for BXT A0/A1 and can be removed when we have newer stable steppings. Durga can add more..

Regards,
Sonika

From: Rodrigo Vivi [mailto:[email protected]]
Sent: Thursday, September 10, 2015 12:54 AM
To: Jindal, Sonika; [email protected]
Subject: Re: [Intel-gfx] [PATCH 6/6] drm/i915/bxt: Fix irq_port for eDP

Nak: I don't believe we need this... Actually I believe we need the opposite... we need to enable HPD on port A for eDP errors handling...

On Fri, Sep 4, 2015 at 6:38 AM Sonika Jindal <[email protected]<mailto:[email protected]>> wrote:
From: Durgadoss R <[email protected]<mailto:[email protected]>>

Currently, HDMI hotplug with eDP as local panel is failing because the HDMI hpd is detected as a long hpd for eDP; and is thus rightfully ignored. But, it should really be handled as an interrupt on port B for HDMI (due to BXT A1 platform having HPD pins A and B swapped). This patch sets the irq_port[PORT_A] to NULL in case eDP is on port A so that irq handler does not treat it as a 'dig_port' interrupt.

---
 drivers/gpu/drm/i915/intel_ddi.c |   15 ++++++++++-----
 1 file changed, 10 insertions(+), 5 deletions(-)
drivers/gpu/drm/i915/intel_ddi.c +10 -5
index 4823184..fec51df 100644
@@ -3218,15 +3218,20 @@ void intel_ddi_init(struct drm_device *dev, enum port port)
                        goto err;
 
                intel_dig_port->hpd_pulse = intel_dp_hpd_pulse;
+               dev_priv->hotplug.irq_port[port] = intel_dig_port;
                /*
                 * On BXT A0/A1, sw needs to activate DDIA HPD logic and
                 * interrupts to check the external panel connection.
+                * If eDP is connected on port A, set irq_port to NULL
+                * so that we do not assume an interrupt here as a
+                * 'dig_port' interrupt.
                 */
-               if (IS_BROXTON(dev_priv) && (INTEL_REVID(dev) < BXT_REVID_B0)
-                                        && port == PORT_B)
-                       dev_priv->hotplug.irq_port[PORT_A] = intel_dig_port;
-               else
-                       dev_priv->hotplug.irq_port[port] = intel_dig_port;
+               if (IS_BROXTON(dev) && (INTEL_REVID(dev) < BXT_REVID_B0)) {
+                       if (port == PORT_B)
+                               dev_priv->hotplug.irq_port[PORT_A] = intel_dig_port;
+                       else if (intel_encoder->type == INTEL_OUTPUT_EDP)
+                               dev_priv->hotplug.irq_port[port] = NULL;
+               }
        }
 
        /* In theory we don't need the encoder->type check, but leave it just in

--
1.7.10.4

_______________________________________________
Intel-gfx mailing list [email protected]<mailto:[email protected]>
http://lists.freedesktop.org/mailman/listinfo/intel-gfx

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

#71 lore raw

Re: [PATCH 6/6] drm/i915/bxt: Fix irq_port for eDP

In reply to Rodrigo Vivi

Hi Rodrigo,

I had to add this to get HDMI hotplug working on BXT. As you might already know, we have the HPD pins A & B swapped in BXT. And, we are using HPD_PORT_A as 'intel_encoder->hpd_pin' for HDMI on port B.

Without this, When an HPD on HDMI (port B) happens, the interrupt is handled as an eDP (port A) interrupt in 'intel_hpd_irq_handler', since hpd_pin for HDMI port B is set as PIN A.

Snippet:
---
is_dig_port = intel_hpd_pin_to_port(i, &port) &&

                               dev_priv->hotplug.irq_port[port];

---

The issue occurs only when we have eDP + HDMI combination. MIPI + HDMI works well without this fix also.

And all these workarounds, are due to pin swap in BXT A0/A1. We have this fix enclosed with that check as well.

I tried to few other changes, but this one was less intrusive and easy to change (and notice the change).
Kindly let us know if you have better ideas.

Thanks,
Durga

From: Intel-gfx [mailto:[email protected]] On Behalf Of Rodrigo Vivi
Sent: Thursday, September 10, 2015 12:54 AM
To: Jindal, Sonika; [email protected]
Subject: Re: [Intel-gfx] [PATCH 6/6] drm/i915/bxt: Fix irq_port for eDP

Nak: I don't believe we need this... Actually I believe we need the opposite... we need to enable HPD on port A for eDP errors handling...

On Fri, Sep 4, 2015 at 6:38 AM Sonika Jindal <[email protected]> wrote:
From: Durgadoss R <[email protected]>

Currently, HDMI hotplug with eDP as local panel is failing because the HDMI hpd is detected as a long hpd for eDP; and is thus rightfully ignored. But, it should really be handled as an interrupt on port B for HDMI (due to BXT A1 platform having HPD pins A and B swapped). This patch sets the irq_port[PORT_A] to NULL in case eDP is on port A so that irq handler does not treat it as a 'dig_port' interrupt.

---
 drivers/gpu/drm/i915/intel_ddi.c |   15 ++++++++++-----

 1 file changed, 10 insertions(+), 5 deletions(-)
drivers/gpu/drm/i915/intel_ddi.c +0 -0
index 4823184..fec51df 100644
@@ -3218,15 +3218,20 @@ void intel_ddi_init(struct drm_device *dev, enum port port)

                        goto err;

                intel_dig_port->hpd_pulse = intel_dp_hpd_pulse;
+               dev_priv->hotplug.irq_port[port] = intel_dig_port;
                /*
                 * On BXT A0/A1, sw needs to activate DDIA HPD logic and
                 * interrupts to check the external panel connection. +                * If eDP is connected on port A, set irq_port to NULL +                * so that we do not assume an interrupt here as a
+                * 'dig_port' interrupt.
                 */
-               if (IS_BROXTON(dev_priv) && (INTEL_REVID(dev) < BXT_REVID_B0) -                                        && port == PORT_B)
-                       dev_priv->hotplug.irq_port[PORT_A] = intel_dig_port; -               else
-                       dev_priv->hotplug.irq_port[port] = intel_dig_port; +               if (IS_BROXTON(dev) && (INTEL_REVID(dev) < BXT_REVID_B0)) { +                       if (port == PORT_B)
+                               dev_priv->hotplug.irq_port[PORT_A] = intel_dig_port; +                       else if (intel_encoder->type == INTEL_OUTPUT_EDP) +                               dev_priv->hotplug.irq_port[port] = NULL; +               }
        }

        /* In theory we don't need the encoder->type check, but leave it just in --
1.7.10.4

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

#72 lore raw

Re: [PATCH 6/6] drm/i915/bxt: Fix irq_port for eDP

Quoted 6 lines · Durgadoss

Hi Rodrigo,

I had to add this to get HDMI hotplug working on BXT. As you might already know, we have the HPD pins A & B swapped in BXT. And, we are using HPD_PORT_A as 'intel_encoder->hpd_pin' for HDMI on port B.

People keep saying that, but according to the spec it's not even true. A and B are not swapped, but instead all the pins are sort of shifted by one position.

HPD A = DDI0 / port B
HPD B = DDI1 / port C
HPD C = DDI2 / port A

If the code actually did the remapping in a way that matches the spec, things would be a lot less confusing.

Quoted 92 lines

Without this, When an HPD on HDMI (port B) happens, the interrupt is handled as an eDP (port A) interrupt in 'intel_hpd_irq_handler', since hpd_pin for HDMI port B is set as PIN A.

Snippet:
---
is_dig_port = intel_hpd_pin_to_port(i, &port) &&

                               dev_priv->hotplug.irq_port[port];

---

The issue occurs only when we have eDP + HDMI combination. MIPI + HDMI works well without this fix also.

And all these workarounds, are due to pin swap in BXT A0/A1. We have this fix enclosed with that check as well.

I tried to few other changes, but this one was less intrusive and easy to change (and notice the change).
Kindly let us know if you have better ideas.

Thanks,
Durga

From: Intel-gfx [mailto:[email protected]] On Behalf Of Rodrigo Vivi
Sent: Thursday, September 10, 2015 12:54 AM
To: Jindal, Sonika; [email protected]
Subject: Re: [Intel-gfx] [PATCH 6/6] drm/i915/bxt: Fix irq_port for eDP

Nak: I don't believe we need this... Actually I believe we need the opposite... we need to enable HPD on port A for eDP errors handling...

On Fri, Sep 4, 2015 at 6:38 AM Sonika Jindal <[email protected]> wrote:
From: Durgadoss R <[email protected]>

Currently, HDMI hotplug with eDP as local panel is failing because the HDMI hpd is detected as a long hpd for eDP; and is thus rightfully ignored. But, it should really be handled as an interrupt on port B for HDMI (due to BXT A1 platform having HPD pins A and B swapped). This patch sets the irq_port[PORT_A] to NULL in case eDP is on port A so that irq handler does not treat it as a 'dig_port' interrupt.

---
 drivers/gpu/drm/i915/intel_ddi.c |   15 ++++++++++-----

 1 file changed, 10 insertions(+), 5 deletions(-)
drivers/gpu/drm/i915/intel_ddi.c +10 -6
index 4823184..fec51df 100644
@@ -3218,15 +3218,20 @@ void intel_ddi_init(struct drm_device *dev, enum port port)
                         goto err;
 
                 intel_dig_port->hpd_pulse = intel_dp_hpd_pulse;
+               dev_priv->hotplug.irq_port[port] = intel_dig_port;
                 /*
                  * On BXT A0/A1, sw needs to activate DDIA HPD logic and
                  * interrupts to check the external panel connection.
+                * If eDP is connected on port A, set irq_port to NULL
+                * so that we do not assume an interrupt here as a
+                * 'dig_port' interrupt.
                  */
-               if (IS_BROXTON(dev_priv) && (INTEL_REVID(dev) < BXT_REVID_B0)
-                                        && port == PORT_B)
-                       dev_priv->hotplug.irq_port[PORT_A] = intel_dig_port;
-               else
-                       dev_priv->hotplug.irq_port[port] = intel_dig_port;
+               if (IS_BROXTON(dev) && (INTEL_REVID(dev) < BXT_REVID_B0)) {
+                       if (port == PORT_B)
+                               dev_priv->hotplug.irq_port[PORT_A] = intel_dig_port;
+                       else if (intel_encoder->type == INTEL_OUTPUT_EDP)
+                               dev_priv->hotplug.irq_port[port] = NULL;
+               }
         }
 
         /* In theory we don't need the encoder->type check, but leave it just in
--

1.7.10.4

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

Signature · 6 lines
Ville Syrjälä
Intel OTC
_______________________________________________
Intel-gfx mailing list
[email protected]
http://lists.freedesktop.org/mailman/listinfo/intel-gfx
#73 lore raw

Re: [PATCH 6/6] drm/i915/bxt: Fix irq_port for eDP

Quoted 12 lines · Ville Syrjälä
Quoted 6 lines · Durgadoss

Hi Rodrigo,

I had to add this to get HDMI hotplug working on BXT. As you might already know, we have the HPD pins A & B swapped in BXT. And, we are using HPD_PORT_A as 'intel_encoder->hpd_pin' for HDMI on port B.

People keep saying that, but according to the spec it's not even true. A and B are not swapped, but instead all the pins are sort of shifted by one position.

Quoted 6 lines

HPD A = DDI0 / port B
HPD B = DDI1 / port C
HPD C = DDI2 / port A

If the code actually did the remapping in a way that matches the spec, things would be a lot less confusing.

And we even have the logic to allow that remapping, so I guess the real job is to figure out where that's not done correctly. Not throwing more hacks on top.
-Daniel

Signature · 7 lines
Daniel Vetter
Software Engineer, Intel Corporation
http://blog.ffwll.ch
_______________________________________________
Intel-gfx mailing list
[email protected]
http://lists.freedesktop.org/mailman/listinfo/intel-gfx