Skip to contentSkip to conversation

drm/msm+ti-sn65dsi86: Fix NO_CONNECTOR fallout

PATCH In review Rob Clark 5 participants 21 messages Aug 11, 2021 – Sep 23, 2021

Subsystems (6)
lore
Share & tools
lore

Rob Clark

author
#0 lore raw

cover drm/msm+ti-sn65dsi86: Fix NO_CONNECTOR fallout

From: Rob Clark <[email protected]>

The first patch fixes breakage in drm-next for the ti eDP bridge (which is used on nearly all the snapdragon windows laptops and chromebooks). The second add drm/msm NO_CONNECTOR support, and the final two add NO_CONNECTOR support to the ti eDP bridge.

Would be nice to get at least the first one into drm-next kinda soonish since at the moment a lot of things are broken.

Rob Clark (4):
drm/bridge: ti-sn65dsi86: Avoid creating multiple connectors drm/msm/dsi: Support NO_CONNECTOR bridges drm/bridge: ti-sn65dsi86: Implement bridge->mode_valid() drm/bridge: ti-sn65dsi86: Add NO_CONNECTOR support

 drivers/gpu/drm/bridge/ti-sn65dsi86.c | 61 ++++++++++++++++++++++-----
 drivers/gpu/drm/msm/Kconfig           |  2 +
 drivers/gpu/drm/msm/dsi/dsi_manager.c | 41 ++++++++++++------
 3 files changed, 81 insertions(+), 23 deletions(-)
Signature · 1 line
2.31.1
author
#1 lore raw

PATCH 1/4 drm/bridge: ti-sn65dsi86: Avoid creating multiple connectors

From: Rob Clark <[email protected]>

If we created our own connector because the driver does not support the NO_CONNECTOR flag, we don't want the downstream bridge to *also* create a connector. And if this driver did pass the NO_CONNECTOR flag (and we supported that mode) this would change nothing.

Fixes 4e5763f03e10 ("drm/bridge: ti-sn65dsi86: Wrap panel with panel-bridge")
---
 drivers/gpu/drm/bridge/ti-sn65dsi86.c | 3 +++
 1 file changed, 3 insertions(+)
drivers/gpu/drm/bridge/ti-sn65dsi86.c +3 -0
index 9bf889302bcc..5d3b30b2f547 100644
@@ -736,6 +736,9 @@ static int ti_sn_bridge_attach(struct drm_bridge *bridge,
 	}
 	pdata->dsi = dsi;
 
+	/* We never want the next bridge to *also* create a connector: */
+	flags |= DRM_BRIDGE_ATTACH_NO_CONNECTOR;
+
 	/* Attach the next bridge */
 	ret = drm_bridge_attach(bridge->encoder, pdata->next_bridge,
 				&pdata->bridge, flags);
Signature · 1 line
2.31.1
Tested-by on v1 1/4
#2 lore raw

Re: [PATCH 1/4] drm/bridge: ti-sn65dsi86: Avoid creating multiple connectors

Quoting Rob Clark (2021-08-11 16:52:47)

Quoted 10 lines

From: Rob Clark <[email protected]>

If we created our own connector because the driver does not support the NO_CONNECTOR flag, we don't want the downstream bridge to *also* create a connector. And if this driver did pass the NO_CONNECTOR flag (and we supported that mode) this would change nothing.

Fixes 4e5763f03e10 ("drm/bridge: ti-sn65dsi86: Wrap panel with panel-bridge")

---

Thanks for saving me the packaging effort.

reviewer Reviewed-by on v1 1/4
#3 lore raw

Re: [PATCH 1/4] drm/bridge: ti-sn65dsi86: Avoid creating multiple connectors

In reply to Rob Clark

Hi Rob,

Thank you for the patch.

Quoted 9 lines · Rob Clark

From: Rob Clark <[email protected]>

If we created our own connector because the driver does not support the NO_CONNECTOR flag, we don't want the downstream bridge to *also* create a connector. And if this driver did pass the NO_CONNECTOR flag (and we supported that mode) this would change nothing.

Fixes 4e5763f03e10 ("drm/bridge: ti-sn65dsi86: Wrap panel with panel-bridge")

Makes complete sense.

Quoted 18 lines
---
 drivers/gpu/drm/bridge/ti-sn65dsi86.c | 3 +++
 1 file changed, 3 insertions(+)
drivers/gpu/drm/bridge/ti-sn65dsi86.c +3 -0
index 9bf889302bcc..5d3b30b2f547 100644
@@ -736,6 +736,9 @@ static int ti_sn_bridge_attach(struct drm_bridge *bridge,
 	}
 	pdata->dsi = dsi;
 
+	/* We never want the next bridge to *also* create a connector: */
+	flags |= DRM_BRIDGE_ATTACH_NO_CONNECTOR;
+
 	/* Attach the next bridge */
 	ret = drm_bridge_attach(bridge->encoder, pdata->next_bridge,
 				&pdata->bridge, flags);
Signature · 3 lines
Regards,

Laurent Pinchart
reviewer Reviewed-by Tested-by on v1 1/4
#4 lore raw

Re: [PATCH 1/4] drm/bridge: ti-sn65dsi86: Avoid creating multiple connectors

In reply to Rob Clark

Hi,

Quoted 8 lines · Rob Clark

From: Rob Clark <[email protected]>

If we created our own connector because the driver does not support the NO_CONNECTOR flag, we don't want the downstream bridge to *also* create a connector. And if this driver did pass the NO_CONNECTOR flag (and we supported that mode) this would change nothing.

Quoted 5 lines
Fixes 4e5763f03e10 ("drm/bridge: ti-sn65dsi86: Wrap panel with panel-bridge")
---
 drivers/gpu/drm/bridge/ti-sn65dsi86.c | 3 +++
 1 file changed, 3 insertions(+)

I'm going to apply this one to drm-misc/drm-misc-next and push since it's a fix and we're before -rc6, then I'll take a look at the later patches in the series.

-Doug

author
#5 lore raw

Re: [PATCH 1/4] drm/bridge: ti-sn65dsi86: Avoid creating multiple connectors

Quoted 18 lines · Doug Anderson

Hi,

Quoted 8 lines · Rob Clark

From: Rob Clark <[email protected]>

If we created our own connector because the driver does not support the NO_CONNECTOR flag, we don't want the downstream bridge to *also* create a connector. And if this driver did pass the NO_CONNECTOR flag (and we supported that mode) this would change nothing.

Quoted 5 lines
Fixes 4e5763f03e10 ("drm/bridge: ti-sn65dsi86: Wrap panel with panel-bridge")
---
 drivers/gpu/drm/bridge/ti-sn65dsi86.c | 3 +++
 1 file changed, 3 insertions(+)
Quoted 7 lines

I'm going to apply this one to drm-misc/drm-misc-next and push since it's a fix and we're before -rc6, then I'll take a look at the later patches in the series.

Thanks.. this is the only one with some urgency, the rest can wait until next cycle. (And the bridge vs msm patches can land independently, I've tested the different possible combinations)

BR,
-R

author
#6 lore raw

PATCH 2/4 drm/msm/dsi: Support NO_CONNECTOR bridges

In reply to Rob Clark

From: Rob Clark <[email protected]>

For now, since we have a mix of bridges which support this flag, which which do *not* support this flag, or work both ways, try it once with NO_CONNECTOR and then fall back to the old way if that doesn't work. Eventually we can drop the fallback path.

---
 drivers/gpu/drm/msm/Kconfig           |  2 ++
 drivers/gpu/drm/msm/dsi/dsi_manager.c | 41 ++++++++++++++++++---------
 2 files changed, 30 insertions(+), 13 deletions(-)
drivers/gpu/drm/msm/Kconfig +2 -0
index e9c6af78b1d7..36e5ba3ccc28 100644
@@ -14,6 +14,8 @@ config DRM_MSM
 	select REGULATOR
 	select DRM_KMS_HELPER
 	select DRM_PANEL
+	select DRM_BRIDGE
+	select DRM_PANEL_BRIDGE
 	select DRM_SCHED
 	select SHMEM
 	select TMPFS
drivers/gpu/drm/msm/dsi/dsi_manager.c +28 -13

64 lines not on this page · raw message

Signature · 1 line
2.31.1
reviewer Reviewed-by on v1 2/4
#7 lore raw

Re: [PATCH 2/4] drm/msm/dsi: Support NO_CONNECTOR bridges

Hi Rob

Thank you for the patch.

Quoted 51 lines · Rob Clark

From: Rob Clark <[email protected]>

For now, since we have a mix of bridges which support this flag, which which do *not* support this flag, or work both ways, try it once with NO_CONNECTOR and then fall back to the old way if that doesn't work. Eventually we can drop the fallback path.

---
 drivers/gpu/drm/msm/Kconfig           |  2 ++
 drivers/gpu/drm/msm/dsi/dsi_manager.c | 41 ++++++++++++++++++---------
 2 files changed, 30 insertions(+), 13 deletions(-)
drivers/gpu/drm/msm/Kconfig +2 -0
index e9c6af78b1d7..36e5ba3ccc28 100644
@@ -14,6 +14,8 @@ config DRM_MSM
 	select REGULATOR
 	select DRM_KMS_HELPER
 	select DRM_PANEL
+	select DRM_BRIDGE
+	select DRM_PANEL_BRIDGE
 	select DRM_SCHED
 	select SHMEM
 	select TMPFS
drivers/gpu/drm/msm/dsi/dsi_manager.c +3 -2
index c41d39f5b7cf..1fd1cf93abbf 100644
@@ -3,6 +3,8 @@
  * Copyright (c) 2015, The Linux Foundation. All rights reserved.
  */
 
+#include "drm/drm_bridge_connector.h"
+
 #include "msm_kms.h"
 #include "dsi.h"
 
@@ -690,8 +692,7 @@ struct drm_connector *msm_dsi_manager_ext_bridge_init(u8 id)
 	struct drm_device *dev = msm_dsi->dev;
 	struct drm_encoder *encoder;
 	struct drm_bridge *int_bridge, *ext_bridge;
-	struct drm_connector *connector;
-	struct list_head *connector_list;
+	int ret;
 
 	int_bridge = msm_dsi->bridge;
 	ext_bridge = msm_dsi->external_bridge =
@@ -699,22 +700,36 @@ struct drm_connector *msm_dsi_manager_ext_bridge_init(u8 id)
 

drivers/gpu/drm/msm/dsi/dsi_manager.c struct drm_connector *msm_dsi_manager_ext_bridge_init(u8 id) · line 704

Quoted 10 lines
drivers/gpu/drm/msm/dsi/dsi_manager.c +1 -6
 	encoder = msm_dsi->encoder;
 
-	/* link the internal dsi bridge to the external bridge */
-	drm_bridge_attach(encoder, ext_bridge, int_bridge, 0);
-
 	/*
-	 * we need the drm_connector created by the external bridge
-	 * driver (or someone else) to feed it to our driver's
-	 * priv->connector[] list, mainly for msm_fbdev_init()
+	 * Try first to create the bridge without it creating it's own

s/it's/its/

drivers/gpu/drm/msm/dsi/dsi_manager.c struct drm_connector *msm_dsi_manager_ext_bridge_init(u8 id) · line 709

Quoted 6 lines
drivers/gpu/drm/msm/dsi/dsi_manager.c +4 -1
+	 * connector.. currently some bridges support this, and others
+	 * do not (and some support both modes)
 	 */
-	connector_list = &dev->mode_config.connector_list;
+	ret = drm_bridge_attach(encoder, ext_bridge, int_bridge,
+			DRM_BRIDGE_ATTACH_NO_CONNECTOR);

Should all this be moved one layer up, to the code that attaches to the mem_dsi->bridge ? I suppose we can start here, but as part of a global move to bridges and DRM_BRIDGE_ATTACH_NO_CONNECTOR, I think the top-level would make more sense in the long term.

If you want to start here,

Quoted 30 lines
drivers/gpu/drm/msm/dsi/dsi_manager.c +20 -4
+	if (ret == -EINVAL) {
+		struct drm_connector *connector;
+		struct list_head *connector_list;
+
+		/* link the internal dsi bridge to the external bridge */
+		drm_bridge_attach(encoder, ext_bridge, int_bridge, 0);
+
+		/*
+		 * we need the drm_connector created by the external bridge
+		 * driver (or someone else) to feed it to our driver's
+		 * priv->connector[] list, mainly for msm_fbdev_init()
+		 */
+		connector_list = &dev->mode_config.connector_list;
+
+		list_for_each_entry(connector, connector_list, head) {
+			if (drm_connector_has_possible_encoder(connector, encoder))
+				return connector;
+		}
 
-	list_for_each_entry(connector, connector_list, head) {
-		if (drm_connector_has_possible_encoder(connector, encoder))
-			return connector;
+		return ERR_PTR(-ENODEV);
 	}
 
-	return ERR_PTR(-ENODEV);
+	return drm_bridge_connector_init(dev, encoder);
 }
 
 void msm_dsi_manager_bridge_destroy(struct drm_bridge *bridge)
Signature · 3 lines
Regards,

Laurent Pinchart
#8 lore raw

Re: [PATCH 2/4] drm/msm/dsi: Support NO_CONNECTOR bridges

In reply to Rob Clark

Hi Rob,

Quoted 6 lines · Rob Clark

From: Rob Clark <[email protected]>

For now, since we have a mix of bridges which support this flag, which which do *not* support this flag, or work both ways, try it once with NO_CONNECTOR and then fall back to the old way if that doesn't work. Eventually we can drop the fallback path.

Which bridges are missing support for the NO_CONNECTOR flags that you are using?

Background for the question is that if you are using one of the bridges missing a conversion then we could do the conversion and maybe you could help with some tests.

We in the above sentence is likely me, so it may take some weeks, but known there is a chance for testing is a good motivator.

Sam

author
#9 lore raw

Re: [PATCH 2/4] drm/msm/dsi: Support NO_CONNECTOR bridges

Quoted 6 lines · Sam Ravnborg

Hi Rob,

Quoted 2 lines · Rob Clark

From: Rob Clark <[email protected]>

Quoted 7 lines
Quoted 4 lines

For now, since we have a mix of bridges which support this flag, which which do *not* support this flag, or work both ways, try it once with NO_CONNECTOR and then fall back to the old way if that doesn't work. Eventually we can drop the fallback path.

Which bridges are missing support for the NO_CONNECTOR flags that you are using?

Currently (as far as things that I am aware of) it is just ti-sn65dsi86, which the last two patches in this series convert.

But who knows what someone somewhere decides to build ;-)

BR,
-R

Quoted 8 lines

Background for the question is that if you are using one of the bridges missing a conversion then we could do the conversion and maybe you could help with some tests.

We in the above sentence is likely me, so it may take some weeks, but known there is a chance for testing is a good motivator.

Sam

author
#10 lore raw

PATCH 3/4 drm/bridge: ti-sn65dsi86: Implement bridge->mode_valid()

In reply to Rob Clark

From: Rob Clark <[email protected]>

For the brave new world of bridges not creating their own connectors, we need to implement the max clock limitation via bridge->mode_valid() instead of connector->mode_valid().

---
 drivers/gpu/drm/bridge/ti-sn65dsi86.c | 24 +++++++++++++++++++-----
 1 file changed, 19 insertions(+), 5 deletions(-)
drivers/gpu/drm/bridge/ti-sn65dsi86.c +19 -5
index 5d3b30b2f547..38dcc49eccaf 100644
@@ -595,6 +595,15 @@ static struct auxiliary_driver ti_sn_aux_driver = {
 	.id_table = ti_sn_aux_id_table,
 };
 
+static enum drm_mode_status check_mode(const struct drm_display_mode *mode)
+{
+	/* maximum supported resolution is 4K at 60 fps */
+	if (mode->clock > 594000)
+		return MODE_CLOCK_HIGH;
+
+	return MODE_OK;
+}
+
 /* -----------------------------------------------------------------------------
  * DRM Connector Operations
  */
@@ -616,11 +625,7 @@ static enum drm_mode_status
 ti_sn_bridge_connector_mode_valid(struct drm_connector *connector,
 				  struct drm_display_mode *mode)
 {
-	/* maximum supported resolution is 4K at 60 fps */
-	if (mode->clock > 594000)
-		return MODE_CLOCK_HIGH;
-
-	return MODE_OK;
+	return check_mode(mode);
 }
 
 static struct drm_connector_helper_funcs ti_sn_bridge_connector_helper_funcs = {
@@ -763,6 +768,14 @@ static void ti_sn_bridge_detach(struct drm_bridge *bridge)
 	drm_dp_aux_unregister(&bridge_to_ti_sn65dsi86(bridge)->aux);
 }
 
+static enum drm_mode_status
+ti_sn_bridge_mode_valid(struct drm_bridge *bridge,
+			const struct drm_display_info *info,
+			const struct drm_display_mode *mode)
+{
+	return check_mode(mode);
+}
+
 static void ti_sn_bridge_disable(struct drm_bridge *bridge)
 {
 	struct ti_sn65dsi86 *pdata = bridge_to_ti_sn65dsi86(bridge);
@@ -1118,6 +1131,7 @@ static void ti_sn_bridge_post_disable(struct drm_bridge *bridge)
 static const struct drm_bridge_funcs ti_sn_bridge_funcs = {
 	.attach = ti_sn_bridge_attach,
 	.detach = ti_sn_bridge_detach,
+	.mode_valid = ti_sn_bridge_mode_valid,
 	.pre_enable = ti_sn_bridge_pre_enable,
 	.enable = ti_sn_bridge_enable,
 	.disable = ti_sn_bridge_disable,
Signature · 1 line
2.31.1
reviewer Reviewed-by on v1 3/4
#11 lore raw

Re: [PATCH 3/4] drm/bridge: ti-sn65dsi86: Implement bridge->mode_valid()

Hi,

Quoted 11 lines · Rob Clark

From: Rob Clark <[email protected]>

For the brave new world of bridges not creating their own connectors, we need to implement the max clock limitation via bridge->mode_valid() instead of connector->mode_valid().

---
 drivers/gpu/drm/bridge/ti-sn65dsi86.c | 24 +++++++++++++++++++-----
 1 file changed, 19 insertions(+), 5 deletions(-)

This looks good to me. I'll plan to land this together with the next patch into drm-misc-next sometime next week unless someone beats me to it.

#12 lore raw

Re: [PATCH 3/4] drm/bridge: ti-sn65dsi86: Implement bridge->mode_valid()

In reply to Rob Clark

Hi Rob,

Thank you for the patch.

Quoted 32 lines · Rob Clark

From: Rob Clark <[email protected]>

For the brave new world of bridges not creating their own connectors, we need to implement the max clock limitation via bridge->mode_valid() instead of connector->mode_valid().

---
 drivers/gpu/drm/bridge/ti-sn65dsi86.c | 24 +++++++++++++++++++-----
 1 file changed, 19 insertions(+), 5 deletions(-)
drivers/gpu/drm/bridge/ti-sn65dsi86.c +9 -0
index 5d3b30b2f547..38dcc49eccaf 100644
@@ -595,6 +595,15 @@ static struct auxiliary_driver ti_sn_aux_driver = {
 	.id_table = ti_sn_aux_id_table,
 };
 
+static enum drm_mode_status check_mode(const struct drm_display_mode *mode)
+{
+	/* maximum supported resolution is 4K at 60 fps */
+	if (mode->clock > 594000)
+		return MODE_CLOCK_HIGH;
+
+	return MODE_OK;
+}
+
 /* -----------------------------------------------------------------------------
  * DRM Connector Operations
  */

drivers/gpu/drm/bridge/ti-sn65dsi86.c static enum drm_mode_status · line 628

Quoted 9 lines
drivers/gpu/drm/bridge/ti-sn65dsi86.c +1 -5
@@ -616,11 +625,7 @@ static enum drm_mode_status
 ti_sn_bridge_connector_mode_valid(struct drm_connector *connector,
 				  struct drm_display_mode *mode)
 {
-	/* maximum supported resolution is 4K at 60 fps */
-	if (mode->clock > 594000)
-		return MODE_CLOCK_HIGH;
-
-	return MODE_OK;
+	return check_mode(mode);

Do we need to implement the connector .mode_valid() operation, given that the bridge is linked in the chain ?

Quoted 26 lines
drivers/gpu/drm/bridge/ti-sn65dsi86.c +9 -0
 }
 
 static struct drm_connector_helper_funcs ti_sn_bridge_connector_helper_funcs = {
@@ -763,6 +768,14 @@ static void ti_sn_bridge_detach(struct drm_bridge *bridge)
 	drm_dp_aux_unregister(&bridge_to_ti_sn65dsi86(bridge)->aux);
 }
 
+static enum drm_mode_status
+ti_sn_bridge_mode_valid(struct drm_bridge *bridge,
+			const struct drm_display_info *info,
+			const struct drm_display_mode *mode)
+{
+	return check_mode(mode);
+}
+
 static void ti_sn_bridge_disable(struct drm_bridge *bridge)
 {
 	struct ti_sn65dsi86 *pdata = bridge_to_ti_sn65dsi86(bridge);
@@ -1118,6 +1131,7 @@ static void ti_sn_bridge_post_disable(struct drm_bridge *bridge)
 static const struct drm_bridge_funcs ti_sn_bridge_funcs = {
 	.attach = ti_sn_bridge_attach,
 	.detach = ti_sn_bridge_detach,
+	.mode_valid = ti_sn_bridge_mode_valid,
 	.pre_enable = ti_sn_bridge_pre_enable,
 	.enable = ti_sn_bridge_enable,
 	.disable = ti_sn_bridge_disable,
Signature · 3 lines
Regards,

Laurent Pinchart
author
#13 lore raw

Re: [PATCH 3/4] drm/bridge: ti-sn65dsi86: Implement bridge->mode_valid()

Quoted 42 lines · Laurent Pinchart

Hi Rob,

Thank you for the patch.

Quoted 36 lines · Rob Clark

From: Rob Clark <[email protected]>

For the brave new world of bridges not creating their own connectors, we need to implement the max clock limitation via bridge->mode_valid() instead of connector->mode_valid().

---
 drivers/gpu/drm/bridge/ti-sn65dsi86.c | 24 +++++++++++++++++++-----
 1 file changed, 19 insertions(+), 5 deletions(-)
drivers/gpu/drm/bridge/ti-sn65dsi86.c +9 -1
index 5d3b30b2f547..38dcc49eccaf 100644
@@ -595,6 +595,15 @@ static struct auxiliary_driver ti_sn_aux_driver = {
      .id_table = ti_sn_aux_id_table,
 };
 
+static enum drm_mode_status check_mode(const struct drm_display_mode *mode)
+{
+     /* maximum supported resolution is 4K at 60 fps */
+     if (mode->clock > 594000)
+             return MODE_CLOCK_HIGH;
+
+     return MODE_OK;
+}
+
 /* -----------------------------------------------------------------------------
  * DRM Connector Operations
  */
@@ -616,11 +625,7 @@ static enum drm_mode_status
 ti_sn_bridge_connector_mode_valid(struct drm_connector *connector,
                                struct drm_display_mode *mode)
 {
-     /* maximum supported resolution is 4K at 60 fps */
Quoted 8 lines

drivers/gpu/drm/bridge/ti-sn65dsi86.c static enum drm_mode_status · line 628

Quoted 5 lines
drivers/gpu/drm/bridge/ti-sn65dsi86.c +1 -4
-     if (mode->clock > 594000)
-             return MODE_CLOCK_HIGH;
-
-     return MODE_OK;
+     return check_mode(mode);

Do we need to implement the connector .mode_valid() operation, given that the bridge is linked in the chain ?

My understanding is that we need to keep it for display drivers that are not converted to NO_CONNECTOR..

But AFAIK snapdragon is the only upstream user of this bridge, so after the drm/msm/dsi patch lands we could probably garbage collect the connector support.

BR,
-R

Quoted 31 lines
Quoted 19 lines
drivers/gpu/drm/bridge/ti-sn65dsi86.c +8 -0
 }
 
 static struct drm_connector_helper_funcs ti_sn_bridge_connector_helper_funcs = {
@@ -763,6 +768,14 @@ static void ti_sn_bridge_detach(struct drm_bridge *bridge)
      drm_dp_aux_unregister(&bridge_to_ti_sn65dsi86(bridge)->aux);
 }
 
+static enum drm_mode_status
+ti_sn_bridge_mode_valid(struct drm_bridge *bridge,
+                     const struct drm_display_info *info,
+                     const struct drm_display_mode *mode)
+{
+     return check_mode(mode);
+}
+
 static void ti_sn_bridge_disable(struct drm_bridge *bridge)
 {
      struct ti_sn65dsi86 *pdata = bridge_to_ti_sn65dsi86(bridge);

drivers/gpu/drm/bridge/ti-sn65dsi86.c static void ti_sn_bridge_post_disable(struct drm_bridge *bridge) · line 1137

Quoted 7 lines
drivers/gpu/drm/bridge/ti-sn65dsi86.c +1 -0
@@ -1118,6 +1131,7 @@ static void ti_sn_bridge_post_disable(struct drm_bridge *bridge)
 static const struct drm_bridge_funcs ti_sn_bridge_funcs = {
      .attach = ti_sn_bridge_attach,
      .detach = ti_sn_bridge_detach,
+     .mode_valid = ti_sn_bridge_mode_valid,
      .pre_enable = ti_sn_bridge_pre_enable,
      .enable = ti_sn_bridge_enable,
      .disable = ti_sn_bridge_disable,

--
Regards,

Laurent Pinchart

#14 lore raw

Re: [PATCH 3/4] drm/bridge: ti-sn65dsi86: Implement bridge->mode_valid()

Hi Rob,

Quoted 47 lines · Rob Clark
Quoted 37 lines · Laurent Pinchart
Quoted 36 lines · Rob Clark

From: Rob Clark <[email protected]>

For the brave new world of bridges not creating their own connectors, we need to implement the max clock limitation via bridge->mode_valid() instead of connector->mode_valid().

---
 drivers/gpu/drm/bridge/ti-sn65dsi86.c | 24 +++++++++++++++++++-----
 1 file changed, 19 insertions(+), 5 deletions(-)
drivers/gpu/drm/bridge/ti-sn65dsi86.c +9 -1
index 5d3b30b2f547..38dcc49eccaf 100644
@@ -595,6 +595,15 @@ static struct auxiliary_driver ti_sn_aux_driver = {
      .id_table = ti_sn_aux_id_table,
 };
 
+static enum drm_mode_status check_mode(const struct drm_display_mode *mode)
+{
+     /* maximum supported resolution is 4K at 60 fps */
+     if (mode->clock > 594000)
+             return MODE_CLOCK_HIGH;
+
+     return MODE_OK;
+}
+
 /* -----------------------------------------------------------------------------
  * DRM Connector Operations
  */
@@ -616,11 +625,7 @@ static enum drm_mode_status
 ti_sn_bridge_connector_mode_valid(struct drm_connector *connector,
                                struct drm_display_mode *mode)
 {
-     /* maximum supported resolution is 4K at 60 fps */
Quoted 8 lines

drivers/gpu/drm/bridge/ti-sn65dsi86.c static enum drm_mode_status · line 628

Quoted 5 lines
drivers/gpu/drm/bridge/ti-sn65dsi86.c +1 -4
-     if (mode->clock > 594000)
-             return MODE_CLOCK_HIGH;
-
-     return MODE_OK;
+     return check_mode(mode);

Do we need to implement the connector .mode_valid() operation, given that the bridge is linked in the chain ?

Quoted 6 lines

My understanding is that we need to keep it for display drivers that are not converted to NO_CONNECTOR..

But AFAIK snapdragon is the only upstream user of this bridge, so after the drm/msm/dsi patch lands we could probably garbage collect the connector support.

Even in the !NO_CONNECTOR case, the bridge will still be linked in the chain, so the atomic helpers should call the bridge .mode_valid() in addition to the connector .mode_valid(). I think the connector operation is redundant.

Quoted 26 lines
Quoted 26 lines
Quoted 26 lines
drivers/gpu/drm/bridge/ti-sn65dsi86.c +9 -0
 }
 
 static struct drm_connector_helper_funcs ti_sn_bridge_connector_helper_funcs = {
@@ -763,6 +768,14 @@ static void ti_sn_bridge_detach(struct drm_bridge *bridge)
      drm_dp_aux_unregister(&bridge_to_ti_sn65dsi86(bridge)->aux);
 }
 
+static enum drm_mode_status
+ti_sn_bridge_mode_valid(struct drm_bridge *bridge,
+                     const struct drm_display_info *info,
+                     const struct drm_display_mode *mode)
+{
+     return check_mode(mode);
+}
+
 static void ti_sn_bridge_disable(struct drm_bridge *bridge)
 {
      struct ti_sn65dsi86 *pdata = bridge_to_ti_sn65dsi86(bridge);
@@ -1118,6 +1131,7 @@ static void ti_sn_bridge_post_disable(struct drm_bridge *bridge)
 static const struct drm_bridge_funcs ti_sn_bridge_funcs = {
      .attach = ti_sn_bridge_attach,
      .detach = ti_sn_bridge_detach,
+     .mode_valid = ti_sn_bridge_mode_valid,
      .pre_enable = ti_sn_bridge_pre_enable,
      .enable = ti_sn_bridge_enable,
      .disable = ti_sn_bridge_disable,
Signature · 3 lines
Regards,

Laurent Pinchart
author
#15 lore raw

PATCH 4/4 drm/bridge: ti-sn65dsi86: Add NO_CONNECTOR support

In reply to Rob Clark

From: Rob Clark <[email protected]>

Slightly awkward to fish out the display_info when we aren't creating own connector. But I don't see an obvious better way.

---
 drivers/gpu/drm/bridge/ti-sn65dsi86.c | 34 +++++++++++++++++++++++----
 1 file changed, 29 insertions(+), 5 deletions(-)
drivers/gpu/drm/bridge/ti-sn65dsi86.c +29 -5
index 38dcc49eccaf..dc8112bab3d3 100644
@@ -693,9 +693,11 @@ static int ti_sn_bridge_attach(struct drm_bridge *bridge,
 		return ret;
 	}
 
-	ret = ti_sn_bridge_connector_init(pdata);
-	if (ret < 0)
-		goto err_conn_init;
+	if (!(flags & DRM_BRIDGE_ATTACH_NO_CONNECTOR)) {
+		ret = ti_sn_bridge_connector_init(pdata);
+		if (ret < 0)
+			goto err_conn_init;
+	}
 
 	/*
 	 * TODO: ideally finding host resource and dsi dev registration needs
@@ -757,7 +759,8 @@ static int ti_sn_bridge_attach(struct drm_bridge *bridge,
 err_dsi_attach:
 	mipi_dsi_device_unregister(dsi);
 err_dsi_host:
-	drm_connector_cleanup(&pdata->connector);
+	if (!(flags & DRM_BRIDGE_ATTACH_NO_CONNECTOR))
+		drm_connector_cleanup(&pdata->connector);
 err_conn_init:
 	drm_dp_aux_unregister(&pdata->aux);
 	return ret;
@@ -806,9 +809,30 @@ static void ti_sn_bridge_set_dsi_rate(struct ti_sn65dsi86 *pdata)
 	regmap_write(pdata->regmap, SN_DSIA_CLK_FREQ_REG, val);
 }
 
+/*
+ * Find the connector and fish out the bpc from display_info.  It would
+ * be nice if we could get this instead from drm_bridge_state, but that
+ * doesn't yet appear to be the case.
+ */
 static unsigned int ti_sn_bridge_get_bpp(struct ti_sn65dsi86 *pdata)
 {
-	if (pdata->connector.display_info.bpc <= 6)
+	struct drm_bridge *bridge = &pdata->bridge;
+	struct drm_connector_list_iter conn_iter;
+	struct drm_connector *connector;
+	unsigned bpc = 0;
+
+	drm_connector_list_iter_begin(bridge->dev, &conn_iter);
+	drm_for_each_connector_iter(connector, &conn_iter) {
+		if (drm_connector_has_possible_encoder(connector, bridge->encoder)) {
+			bpc = connector->display_info.bpc;
+			break;
+		}
+	}
+	drm_connector_list_iter_end(&conn_iter);
+
+	WARN_ON(bpc == 0);
+
+	if (bpc <= 6)
 		return 18;
 	else
 		return 24;
Signature · 1 line
2.31.1
reviewer Reviewed-by on v1 4/4
#16 lore raw

Re: [PATCH 4/4] drm/bridge: ti-sn65dsi86: Add NO_CONNECTOR support

Hi,

Quoted 55 lines · Rob Clark

From: Rob Clark <[email protected]>

Slightly awkward to fish out the display_info when we aren't creating own connector. But I don't see an obvious better way.

---
 drivers/gpu/drm/bridge/ti-sn65dsi86.c | 34 +++++++++++++++++++++++----
 1 file changed, 29 insertions(+), 5 deletions(-)
drivers/gpu/drm/bridge/ti-sn65dsi86.c +15 -5
index 38dcc49eccaf..dc8112bab3d3 100644
@@ -693,9 +693,11 @@ static int ti_sn_bridge_attach(struct drm_bridge *bridge,
                return ret;
        }
 
-       ret = ti_sn_bridge_connector_init(pdata);
-       if (ret < 0)
-               goto err_conn_init;
+       if (!(flags & DRM_BRIDGE_ATTACH_NO_CONNECTOR)) {
+               ret = ti_sn_bridge_connector_init(pdata);
+               if (ret < 0)
+                       goto err_conn_init;
+       }
 
        /*
         * TODO: ideally finding host resource and dsi dev registration needs
@@ -757,7 +759,8 @@ static int ti_sn_bridge_attach(struct drm_bridge *bridge,
 err_dsi_attach:
        mipi_dsi_device_unregister(dsi);
 err_dsi_host:
-       drm_connector_cleanup(&pdata->connector);
+       if (!(flags & DRM_BRIDGE_ATTACH_NO_CONNECTOR))
+               drm_connector_cleanup(&pdata->connector);
 err_conn_init:
        drm_dp_aux_unregister(&pdata->aux);
        return ret;
@@ -806,9 +809,30 @@ static void ti_sn_bridge_set_dsi_rate(struct ti_sn65dsi86 *pdata)
        regmap_write(pdata->regmap, SN_DSIA_CLK_FREQ_REG, val);
 }
 
+/*
+ * Find the connector and fish out the bpc from display_info.  It would
+ * be nice if we could get this instead from drm_bridge_state, but that
+ * doesn't yet appear to be the case.
+ */
 static unsigned int ti_sn_bridge_get_bpp(struct ti_sn65dsi86 *pdata)
 {
-       if (pdata->connector.display_info.bpc <= 6)
+       struct drm_bridge *bridge = &pdata->bridge;
+       struct drm_connector_list_iter conn_iter;
+       struct drm_connector *connector;

drivers/gpu/drm/bridge/ti-sn65dsi86.c static void ti_sn_bridge_set_dsi_rate(struct ti_sn65dsi86 *pdata) · line 831

Quoted 10 lines
drivers/gpu/drm/bridge/ti-sn65dsi86.c +10 -0
+       unsigned bpc = 0;
+
+       drm_connector_list_iter_begin(bridge->dev, &conn_iter);
+       drm_for_each_connector_iter(connector, &conn_iter) {
+               if (drm_connector_has_possible_encoder(connector, bridge->encoder)) {
+                       bpc = connector->display_info.bpc;
+                       break;
+               }
+       }
+       drm_connector_list_iter_end(&conn_iter);

This looks reasonable to me. I'll plan to apply it to drm-misc-next sometime next week to give Laurent a chance to comment on whether this causes any problems with his planned support for full DP using this bridge chip. IIUC that means it'll hit mainline 1 rev later, but as per IRC comments this should be fine.

-Doug

#17 lore raw

Re: [PATCH 4/4] drm/bridge: ti-sn65dsi86: Add NO_CONNECTOR support

In reply to Rob Clark

Hi Rob,

Thank you for the patch.

Quoted 4 lines · Rob Clark

From: Rob Clark <[email protected]>

Slightly awkward to fish out the display_info when we aren't creating own connector. But I don't see an obvious better way.

We need a bit more than this, to support the NO_CONNECTOR case, the bridge has to implement a few extra operations, and set the bridge .ops field. I've submitted two patches to do so a while ago:

- [RFC PATCH 08/11] drm/bridge: ti-sn65dsi86: Implement bridge connector operations ([1])
- [RFC PATCH 09/11] drm/bridge: ti-sn65dsi86: Make connector creation optional ([2])

The second patch is similar to the first half of this patch, but misses the cleanup code. I'll try to rebase this and resubmit, but it may take a bit of time.

[1] https://lore.kernel.org/dri-devel/[email protected]/
[2] https://lore.kernel.org/dri-devel/[email protected]/

Quoted 35 lines
---
 drivers/gpu/drm/bridge/ti-sn65dsi86.c | 34 +++++++++++++++++++++++----
 1 file changed, 29 insertions(+), 5 deletions(-)
drivers/gpu/drm/bridge/ti-sn65dsi86.c +7 -4
index 38dcc49eccaf..dc8112bab3d3 100644
@@ -693,9 +693,11 @@ static int ti_sn_bridge_attach(struct drm_bridge *bridge,
 		return ret;
 	}
 
-	ret = ti_sn_bridge_connector_init(pdata);
-	if (ret < 0)
-		goto err_conn_init;
+	if (!(flags & DRM_BRIDGE_ATTACH_NO_CONNECTOR)) {
+		ret = ti_sn_bridge_connector_init(pdata);
+		if (ret < 0)
+			goto err_conn_init;
+	}
 
 	/*
 	 * TODO: ideally finding host resource and dsi dev registration needs
@@ -757,7 +759,8 @@ static int ti_sn_bridge_attach(struct drm_bridge *bridge,
 err_dsi_attach:
 	mipi_dsi_device_unregister(dsi);
 err_dsi_host:
-	drm_connector_cleanup(&pdata->connector);
+	if (!(flags & DRM_BRIDGE_ATTACH_NO_CONNECTOR))
+		drm_connector_cleanup(&pdata->connector);
 err_conn_init:
 	drm_dp_aux_unregister(&pdata->aux);
 	return ret;

drivers/gpu/drm/bridge/ti-sn65dsi86.c static void ti_sn_bridge_set_dsi_rate(struct ti_sn65dsi86 *pdata) · line 816

Quoted 8 lines
drivers/gpu/drm/bridge/ti-sn65dsi86.c +5 -0
@@ -806,9 +809,30 @@ static void ti_sn_bridge_set_dsi_rate(struct ti_sn65dsi86 *pdata)
 	regmap_write(pdata->regmap, SN_DSIA_CLK_FREQ_REG, val);
 }
 
+/*
+ * Find the connector and fish out the bpc from display_info.  It would
+ * be nice if we could get this instead from drm_bridge_state, but that
+ * doesn't yet appear to be the case.
+ */

This should be done with

	struct drm_atomic_state *state = old_bridge_state->base.state;
	struct drm_connector *connector;

	connector = drm_atomic_get_new_connector_for_encoder(state,
							     bridge->encoder);
Quoted 23 lines
drivers/gpu/drm/bridge/ti-sn65dsi86.c +17 -1
 static unsigned int ti_sn_bridge_get_bpp(struct ti_sn65dsi86 *pdata)
 {
-	if (pdata->connector.display_info.bpc <= 6)
+	struct drm_bridge *bridge = &pdata->bridge;
+	struct drm_connector_list_iter conn_iter;
+	struct drm_connector *connector;
+	unsigned bpc = 0;
+
+	drm_connector_list_iter_begin(bridge->dev, &conn_iter);
+	drm_for_each_connector_iter(connector, &conn_iter) {
+		if (drm_connector_has_possible_encoder(connector, bridge->encoder)) {
+			bpc = connector->display_info.bpc;
+			break;
+		}
+	}
+	drm_connector_list_iter_end(&conn_iter);
+
+	WARN_ON(bpc == 0);
+
+	if (bpc <= 6)
 		return 18;
 	else
 		return 24;
Signature · 3 lines
Regards,

Laurent Pinchart
reviewer
#18 lore raw

Re: [PATCH 4/4] drm/bridge: ti-sn65dsi86: Add NO_CONNECTOR support

Laurent,

Quoted 8 lines · Laurent Pinchart

Hi Rob,

Thank you for the patch.

Quoted 2 lines · Rob Clark

From: Rob Clark <[email protected]>

Quoted 8 lines
Quoted 2 lines

Slightly awkward to fish out the display_info when we aren't creating own connector. But I don't see an obvious better way.

We need a bit more than this, to support the NO_CONNECTOR case, the bridge has to implement a few extra operations, and set the bridge .ops field. I've submitted two patches to do so a while ago:

- [RFC PATCH 08/11] drm/bridge: ti-sn65dsi86: Implement bridge connector operations ([1])

Rob asked me about this over IRC, so if he left it out and it's needed then it's my fault. However, I don't believe it's needed until your series making this bridge chip support full DP. For the the eDP case the bridge chip driver in ToT no longer queries the EDID itself. It simply provides an AUX bus to the panel driver and the panel driver queries the EDID. I think that means we don't need to add DRM_BRIDGE_OP_EDID, right?

I was also wondering if in the full DP case we should actually model the physical DP jack as a drm_bridge and have it work the same way. It would get probed via the DP AUX bus just like a panel. I seem to remember Stephen Boyd was talking about modeling the DP connector as a drm_bridge because it would allow us to handle the fact that some TCPC chips could only support HBR2 whereas others could support HBR3. Maybe it would end up being a fairly elegant solution?

Quoted 5 lines

- [RFC PATCH 09/11] drm/bridge: ti-sn65dsi86: Make connector creation optional ([2])

The second patch is similar to the first half of this patch, but misses the cleanup code. I'll try to rebase this and resubmit, but it may take a bit of time.

Whoops! You're right that Rob's patch won't work at all because we'll just hit the "Fix bridge driver to make connector optional!" case. I should have noticed that. :(

-Doug

author
#19 lore raw

Re: [PATCH 4/4] drm/bridge: ti-sn65dsi86: Add NO_CONNECTOR support

Quoted 40 lines · Doug Anderson

Laurent,

Quoted 8 lines · Laurent Pinchart

Hi Rob,

Thank you for the patch.

Quoted 2 lines · Rob Clark

From: Rob Clark <[email protected]>

Quoted 8 lines
Quoted 2 lines

Slightly awkward to fish out the display_info when we aren't creating own connector. But I don't see an obvious better way.

We need a bit more than this, to support the NO_CONNECTOR case, the bridge has to implement a few extra operations, and set the bridge .ops field. I've submitted two patches to do so a while ago:

- [RFC PATCH 08/11] drm/bridge: ti-sn65dsi86: Implement bridge connector operations ([1])

Rob asked me about this over IRC, so if he left it out and it's needed then it's my fault. However, I don't believe it's needed until your series making this bridge chip support full DP. For the the eDP case the bridge chip driver in ToT no longer queries the EDID itself. It simply provides an AUX bus to the panel driver and the panel driver queries the EDID. I think that means we don't need to add DRM_BRIDGE_OP_EDID, right?

I was also wondering if in the full DP case we should actually model the physical DP jack as a drm_bridge and have it work the same way. It would get probed via the DP AUX bus just like a panel. I seem to remember Stephen Boyd was talking about modeling the DP connector as a drm_bridge because it would allow us to handle the fact that some TCPC chips could only support HBR2 whereas others could support HBR3. Maybe it would end up being a fairly elegant solution?

Quoted 2 lines

- [RFC PATCH 09/11] drm/bridge: ti-sn65dsi86: Make connector creation optional ([2])

Quoted 7 lines
Quoted 3 lines

The second patch is similar to the first half of this patch, but misses the cleanup code. I'll try to rebase this and resubmit, but it may take a bit of time.

Whoops! You're right that Rob's patch won't work at all because we'll just hit the "Fix bridge driver to make connector optional!" case. I should have noticed that. :(

Yes, indeed.. once I fix that, I get no display..

Not sure if Laurent is still working on his series, otherwise I can try to figure out what bridge ops are missing

BR,
-R

#20 lore raw

Re: [PATCH 4/4] drm/bridge: ti-sn65dsi86: Add NO_CONNECTOR support

Hi Rob and Doug,

Quoted 14 lines · Rob Clark
Quoted 13 lines · Doug Anderson
Quoted 11 lines · Laurent Pinchart
Quoted 4 lines · Rob Clark

From: Rob Clark <[email protected]>

Slightly awkward to fish out the display_info when we aren't creating own connector. But I don't see an obvious better way.

We need a bit more than this, to support the NO_CONNECTOR case, the bridge has to implement a few extra operations, and set the bridge .ops field. I've submitted two patches to do so a while ago:

- [RFC PATCH 08/11] drm/bridge: ti-sn65dsi86: Implement bridge connector operations ([1])

Quoted 7 lines
Quoted 7 lines

Rob asked me about this over IRC, so if he left it out and it's needed then it's my fault. However, I don't believe it's needed until your series making this bridge chip support full DP. For the the eDP case the bridge chip driver in ToT no longer queries the EDID itself. It simply provides an AUX bus to the panel driver and the panel driver queries the EDID. I think that means we don't need to add DRM_BRIDGE_OP_EDID, right?

That's right.

Quoted 7 lines
Quoted 7 lines

I was also wondering if in the full DP case we should actually model the physical DP jack as a drm_bridge and have it work the same way. It would get probed via the DP AUX bus just like a panel. I seem to remember Stephen Boyd was talking about modeling the DP connector as a drm_bridge because it would allow us to handle the fact that some TCPC chips could only support HBR2 whereas others could support HBR3. Maybe it would end up being a fairly elegant solution?

Physical connectors are already handled as bridges, see drivers/gpu/drm/bridge/display-connector.c. I however don't think it should handle EDID retrieval, because that's really not an operation implemented by the connector itself.

Quoted 6 lines
Quoted 6 lines
Quoted 5 lines

- [RFC PATCH 09/11] drm/bridge: ti-sn65dsi86: Make connector creation optional ([2])

The second patch is similar to the first half of this patch, but misses the cleanup code. I'll try to rebase this and resubmit, but it may take a bit of time.

Quoted 8 lines
Quoted 3 lines

Whoops! You're right that Rob's patch won't work at all because we'll just hit the "Fix bridge driver to make connector optional!" case. I should have noticed that. :(

Yes, indeed.. once I fix that, I get no display..

Not sure if Laurent is still working on his series, otherwise I can try to figure out what bridge ops are missing

I am, but too slowly. I don't mind fast-tracking the changes you need though.

Signature · 3 lines
Regards,

Laurent Pinchart