dt-bindings: mipi-dsi: dual-channel DSI bindings

Add a section that describes dt-bindings for peripherals that support MIPI DSI, but have a different bus as the primary control bus. Add an example for such peripherals.

Signed-off-by Archit Taneja <[email protected]>

---
 .../devicetree/bindings/display/mipi-dsi-bus.txt   | 75 ++++++++++++++++++++--
 1 file changed, 68 insertions(+), 7 deletions(-)

Documentation/devicetree/bindings/display/mipi-dsi-bus.txt +2 -2

index 973c27273772..77a7cec15f5b 100644
@@ -16,7 +16,7 @@ The following assumes that only a single peripheral is connected to a DSI
 host. Experience shows that this is true for the large majority of setups.
 
 DSI host
---------
+========
 
 In addition to the standard properties and those defined by the parent bus of
 a DSI host, the following properties apply to a node representing a DSI host.

@@ -30,11 +30,15 @@ Required properties:
   different value here. See below.
 
 DSI peripheral
---------------
+==============
 

Documentation/devicetree/bindings/display/mipi-dsi-bus.txt +26 -2

 Required properties:
 - reg: The virtual channel number of a DSI peripheral. Must be in the range

@@ -49,9 +53,25 @@ case two alternative representations can be chosen:
   property is the number of the first virtual channel and the second cell is
   the number of consecutive virtual channels.
 
-Example
--------
+Peripherals with a different control bus
+----------------------------------------
+
+There are peripherals that have I2C/SPI (or some other non-DSI bus) as the
+primary control bus, but are also connected to a DSI bus (mostly for the data
+path). Connections between such peripherals and a DSI host can be represented
+using the graph bindings [1], [2].
+
+[1] Documentation/devicetree/bindings/graph.txt
+[2] Documentation/devicetree/bindings/media/video-interfaces.txt
 
+Examples
+========
+- (1), (2) and (3) are examples of a DSI host and peripheral on the DSI bus
+  with different virtual channel configurations.
+- (4) is an example of a peripheral on a I2C control bus connected with to
+  a DSI host using of-graph bindings.
+
+1)
 	dsi-host {
...

@@ -67,6 +87,7 @@ Example
...
 	};
 
+2)
 	dsi-host {
...

@@ -82,6 +103,7 @@ Example
...
 	};
 
+3)
 	dsi-host {
...

@@ -96,3 +118,42 @@ Example
 
...
 	};
+
+4)
+	i2c-host {
+		...
+
+		dsi-bridge@35 {

Documentation/devicetree/bindings/display/mipi-dsi-bus.txt +7 -0

+					bridge_mipi_in: endpoint {
+						remote-endpoint = <&host_mipi_out>;
+					};
+				};
+			};
+		};
+	};

Documentation/devicetree/bindings/display/mipi-dsi-bus.txt +6 -0

+				host_mipi_out: endpoint {
+					remote-endpoint = <&bridge_mipi_in>;
+				};
+			};
+		};
+	};

Signature · 2 lines

The Qualcomm Innovation Center, Inc. is a member of the Code Aurora Forum,
hosted by The Linux Foundation

Quoted 29 lines · Archit Taneja

Add a section that describes dt-bindings for peripherals that support MIPI DSI, but have a different bus as the primary control bus. Add an example for such peripherals.

Signed-off-by Archit Taneja <[email protected]>

---
  .../devicetree/bindings/display/mipi-dsi-bus.txt   | 75 ++++++++++++++++++++--
  1 file changed, 68 insertions(+), 7 deletions(-)

Documentation/devicetree/bindings/display/mipi-dsi-bus.txt +2 -2

index 973c27273772..77a7cec15f5b 100644
@@ -16,7 +16,7 @@ The following assumes that only a single peripheral is connected to a DSI
  host. Experience shows that this is true for the large majority of setups.
 
  DSI host
---------
+========
 
  In addition to the standard properties and those defined by the parent bus of
  a DSI host, the following properties apply to a node representing a DSI host.

@@ -30,11 +30,15 @@ Required properties:
    different value here. See below.
 
  DSI peripheral
---------------
+==============
 

Documentation/devicetree/bindings/display/mipi-dsi-bus.txt Required properties: · line 41

Quoted 10 lines

Documentation/devicetree/bindings/display/mipi-dsi-bus.txt +7 -3

-Peripherals are represented as child nodes of the DSI host's node. Properties
-described here apply to all DSI peripherals, but individual bindings may want
-to define additional, device-specific properties.
+Peripherals with DSI as control bus
+------------------------------------
+
+Peripherals with the DSI bus as the primary control path are represented as
+child nodes of the DSI host's node. Properties described here apply to all DSI
+peripherals, but individual bindings may want to define additional,
+device-specific properties.

Are there any panels with no control bus? I've never seen one, but it should be possible if LVDS panels can power on without commands.

Quoted 57 lines

Documentation/devicetree/bindings/display/mipi-dsi-bus.txt +26 -2

  Required properties:
  - reg: The virtual channel number of a DSI peripheral. Must be in the range

@@ -49,9 +53,25 @@ case two alternative representations can be chosen:
    property is the number of the first virtual channel and the second cell is
    the number of consecutive virtual channels.
 
-Example
--------
+Peripherals with a different control bus
+----------------------------------------
+
+There are peripherals that have I2C/SPI (or some other non-DSI bus) as the
+primary control bus, but are also connected to a DSI bus (mostly for the data
+path). Connections between such peripherals and a DSI host can be represented
+using the graph bindings [1], [2].
+
+[1] Documentation/devicetree/bindings/graph.txt
+[2] Documentation/devicetree/bindings/media/video-interfaces.txt
 
+Examples
+========
+- (1), (2) and (3) are examples of a DSI host and peripheral on the DSI bus
+  with different virtual channel configurations.
+- (4) is an example of a peripheral on a I2C control bus connected with to
+  a DSI host using of-graph bindings.
+
+1)
  	dsi-host {
...

@@ -67,6 +87,7 @@ Example
...
  	};
 
+2)
  	dsi-host {
...

@@ -82,6 +103,7 @@ Example
...
  	};
 
+3)
  	dsi-host {
...

@@ -96,3 +118,42 @@ Example
 
...
  	};
+
+4)
+	i2c-host {
+		...
+
+		dsi-bridge@35 {

Documentation/devicetree/bindings/display/mipi-dsi-bus.txt Example · line 136

Quoted 10 lines

Documentation/devicetree/bindings/display/mipi-dsi-bus.txt +10 -0

+			compatible = "...";
+			reg = <0x35>;
+
+			ports {
+				#address-cells = <1>;
+				#size-cells = <0>;
+
+				...
+
+				port@0 {

Drop unit-address and #*-cells.

Quoted 11 lines

Documentation/devicetree/bindings/display/mipi-dsi-bus.txt +11 -0

+					bridge_mipi_in: endpoint {
+						remote-endpoint = <&host_mipi_out>;
+					};
+				};
+			};
+		};
+	};
+
+	dsi-host {
+		...
+

Documentation/devicetree/bindings/display/mipi-dsi-bus.txt Example · line 136

Quoted 6 lines

Documentation/devicetree/bindings/display/mipi-dsi-bus.txt +6 -0

+		ports {
+			#address-cells = <1>;
+			#size-cells = <0>;
+			...
+
+			port@0 {

Drop unit-address and #*-cells.

Add binding info for peripherals that support dual-channel DSI. Add corresponding optional bindings for DSI host controllers that may be configured in this mode.

Signed-off-by Archit Taneja <[email protected]>

---
 .../devicetree/bindings/display/mipi-dsi-bus.txt   | 77 ++++++++++++++++++++++
 1 file changed, 77 insertions(+)

Documentation/devicetree/bindings/display/mipi-dsi-bus.txt +0 -0

index 77a7cec15f5b..f556aaafdf22 100644

Documentation/devicetree/bindings/display/mipi-dsi-bus.txt +15 -0

+
 DSI peripheral
 ==============
 

@@ -61,6 +67,17 @@ primary control bus, but are also connected to a DSI bus (mostly for the data
 path). Connections between such peripherals and a DSI host can be represented
 using the graph bindings [1], [2].
 
+Peripherals that support dual channel DSI
+-----------------------------------------
+
+Peripherals with higher bandwidth requirements can be connected to 2 DSI
+busses. Each DSI bus/channel drives some portion of the pixel data (generally
+left/right half of each line of the display, or even/odd lines of the display).
+The graph bindings should be used to represent the multiple DSI busses that are
+connected to this peripheral. Each DSI host's output endpoint can be linked to
+an input endpoint of the DSI peripheral.
+
+
 [1] Documentation/devicetree/bindings/graph.txt
 [2] Documentation/devicetree/bindings/media/video-interfaces.txt
 

@@ -70,6 +87,8 @@ Examples
   with different virtual channel configurations.
 - (4) is an example of a peripheral on a I2C control bus connected with to
   a DSI host using of-graph bindings.
+- (5) is an example of 2 DSI hosts driving a dual-channel DSI peripheral,
+  which uses I2C as its primary control bus.
 
 1)
 	dsi-host {

@@ -157,3 +176,61 @@ Examples
 			};
 		};
 	};
+

Documentation/devicetree/bindings/display/mipi-dsi-bus.txt +17 -0

+					dsi0_in: endpoint {
+						remote-endpoint = <&dsi0_out>;
+					};
+				};
+
+				port@1 {
+					dsi1_in: endpoint {
+						remote-endpoint = <&dsi1_out>;
+					};
+				};
+			};
+		};
+	};
+
+	dsi0-host {
+		...
+

Documentation/devicetree/bindings/display/mipi-dsi-bus.txt +4 -0

+				dsi0_out: endpoint {
+					remote-endpoint = <&dsi0_in>;
+				};
+			};

Documentation/devicetree/bindings/display/mipi-dsi-bus.txt +6 -0

+				dsi1_out: endpoint {
+					remote-endpoint = <&dsi1_in>;
+				};
+			};
+		};
+	};

Signature · 2 lines

The Qualcomm Innovation Center, Inc. is a member of the Code Aurora Forum,
hosted by The Linux Foundation

Add a section that describes dt-bindings for peripherals that support MIPI DSI, but have a different bus as the primary control bus, or no control bus at all. Add an example for a peripheral with a non-DSI control bus.

Signed-off-by Archit Taneja <[email protected]>

---
v2:
- Mentioned what to do if peripheral has no control bus
- Drop unit-address and #*-cells where applicable.

 .../devicetree/bindings/display/mipi-dsi-bus.txt   | 71 +++++++++++++++++++---
 1 file changed, 64 insertions(+), 7 deletions(-)

Documentation/devicetree/bindings/display/mipi-dsi-bus.txt +54 -7

index 973c27273772..94fb72cb916f 100644
@@ -16,7 +16,7 @@ The following assumes that only a single peripheral is connected to a DSI
 host. Experience shows that this is true for the large majority of setups.
 
 DSI host
---------
+========
 
 In addition to the standard properties and those defined by the parent bus of
 a DSI host, the following properties apply to a node representing a DSI host.

@@ -30,11 +30,16 @@ Required properties:
   different value here. See below.
 
 DSI peripheral
---------------
+==============
 
-Peripherals are represented as child nodes of the DSI host's node. Properties
-described here apply to all DSI peripherals, but individual bindings may want
-to define additional, device-specific properties.
+Peripherals with DSI as control bus, or no control bus
+------------------------------------------------------
+
+Peripherals with the DSI bus as the primary control bus, or peripherals with
+no control bus but use the DSI bus to transmit pixel data are represented
+as child nodes of the DSI host's node. Properties described here apply to all
+DSI peripherals, but individual bindings may want to define additional,
+device-specific properties.
 
 Required properties:
 - reg: The virtual channel number of a DSI peripheral. Must be in the range

@@ -49,9 +54,25 @@ case two alternative representations can be chosen:
   property is the number of the first virtual channel and the second cell is
   the number of consecutive virtual channels.
 
-Example
--------
+Peripherals with a different control bus
+----------------------------------------
+
+There are peripherals that have I2C/SPI (or some other non-DSI bus) as the
+primary control bus, but are also connected to a DSI bus (mostly for the data
+path). Connections between such peripherals and a DSI host can be represented
+using the graph bindings [1], [2].
+
+[1] Documentation/devicetree/bindings/graph.txt
+[2] Documentation/devicetree/bindings/media/video-interfaces.txt
 
+Examples
+========
+- (1), (2) and (3) are examples of a DSI host and peripheral on the DSI bus
+  with different virtual channel configurations.
+- (4) is an example of a peripheral on a I2C control bus connected with to
+  a DSI host using of-graph bindings.
+
+1)
 	dsi-host {
...

@@ -67,6 +88,7 @@ Example
...
 	};
 
+2)
 	dsi-host {
...

@@ -82,6 +104,7 @@ Example
...
 	};
 
+3)
 	dsi-host {
...

@@ -96,3 +119,37 @@ Example
 
...
 	};
+
+4)
+	i2c-host {
+		...
+
+		dsi-bridge@35 {
+			compatible = "...";
+			reg = <0x35>;
+
+			ports {
+				...
+
+				port {
+					bridge_mipi_in: endpoint {
+						remote-endpoint = <&host_mipi_out>;
+					};
+				};
+			};
+		};
+	};
+
+	dsi-host {
+		...
+

Add a section that describes dt-bindings for peripherals that support MIPI DSI, but have a different bus as the primary control bus, or no control bus at all. Add an example for a peripheral with a non-DSI control bus.

Signed-off-by Archit Taneja <[email protected]>

---
v2:
- Mentioned what to do if peripheral has no control bus
- Drop unit-address and #*-cells where applicable.

  .../devicetree/bindings/display/mipi-dsi-bus.txt   | 71 +++++++++++++++++++---
  1 file changed, 64 insertions(+), 7 deletions(-)

Documentation/devicetree/bindings/display/mipi-dsi-bus.txt +54 -7

index 973c27273772..94fb72cb916f 100644
@@ -16,7 +16,7 @@ The following assumes that only a single peripheral is connected to a DSI
  host. Experience shows that this is true for the large majority of setups.
 
  DSI host
---------
+========
 
  In addition to the standard properties and those defined by the parent bus of
  a DSI host, the following properties apply to a node representing a DSI host.

@@ -30,11 +30,16 @@ Required properties:
    different value here. See below.
 
  DSI peripheral
---------------
+==============
 
-Peripherals are represented as child nodes of the DSI host's node. Properties
-described here apply to all DSI peripherals, but individual bindings may want
-to define additional, device-specific properties.
+Peripherals with DSI as control bus, or no control bus
+------------------------------------------------------
+
+Peripherals with the DSI bus as the primary control bus, or peripherals with
+no control bus but use the DSI bus to transmit pixel data are represented
+as child nodes of the DSI host's node. Properties described here apply to all
+DSI peripherals, but individual bindings may want to define additional,
+device-specific properties.
 
  Required properties:
  - reg: The virtual channel number of a DSI peripheral. Must be in the range

@@ -49,9 +54,25 @@ case two alternative representations can be chosen:
    property is the number of the first virtual channel and the second cell is
    the number of consecutive virtual channels.
 
-Example
--------
+Peripherals with a different control bus
+----------------------------------------
+
+There are peripherals that have I2C/SPI (or some other non-DSI bus) as the
+primary control bus, but are also connected to a DSI bus (mostly for the data
+path). Connections between such peripherals and a DSI host can be represented
+using the graph bindings [1], [2].
+
+[1] Documentation/devicetree/bindings/graph.txt
+[2] Documentation/devicetree/bindings/media/video-interfaces.txt
 
+Examples
+========
+- (1), (2) and (3) are examples of a DSI host and peripheral on the DSI bus
+  with different virtual channel configurations.
+- (4) is an example of a peripheral on a I2C control bus connected with to
+  a DSI host using of-graph bindings.
+
+1)
  	dsi-host {
...

@@ -67,6 +88,7 @@ Example
...
  	};
 
+2)
  	dsi-host {
...

@@ -82,6 +104,7 @@ Example
...
  	};
 
+3)
  	dsi-host {
...

@@ -96,3 +119,37 @@ Example
 
...
  	};
+
+4)
+	i2c-host {
+		...
+
+		dsi-bridge@35 {
+			compatible = "...";
+			reg = <0x35>;
+
+			ports {
+				...
+
+				port {
+					bridge_mipi_in: endpoint {
+						remote-endpoint = <&host_mipi_out>;
+					};
+				};
+			};
+		};
+	};
+
+	dsi-host {
+		...
+

Add a section that describes dt-bindings for peripherals that support MIPI DSI, but have a different bus as the primary control bus, or no control bus at all. Add an example for a peripheral with a non-DSI control bus.

Signed-off-by Archit Taneja <[email protected]>

---
v2:
- Mentioned what to do if peripheral has no control bus
- Drop unit-address and #*-cells where applicable.

 .../devicetree/bindings/display/mipi-dsi-bus.txt   | 71 +++++++++++++++++++---
 1 file changed, 64 insertions(+), 7 deletions(-)

Add a section that describes dt-bindings for peripherals that support MIPI DSI, but have a different bus as the primary control bus, or no control bus at all. Add an example for a peripheral with a non-DSI control bus.

Signed-off-by Archit Taneja <[email protected]>

---
v2:
- Mentioned what to do if peripheral has no control bus
- Drop unit-address and #*-cells where applicable.

 .../devicetree/bindings/display/mipi-dsi-bus.txt   | 71 +++++++++++++++++++---
 1 file changed, 64 insertions(+), 7 deletions(-)

Documentation/devicetree/bindings/display/mipi-dsi-bus.txt +16 -7

index 973c27273772..94fb72cb916f 100644
@@ -16,7 +16,7 @@ The following assumes that only a single peripheral is connected to a DSI
 host. Experience shows that this is true for the large majority of setups.
 
 DSI host
---------
+========
 
 In addition to the standard properties and those defined by the parent bus of
 a DSI host, the following properties apply to a node representing a DSI host.

@@ -30,11 +30,16 @@ Required properties:
   different value here. See below.
 
 DSI peripheral
---------------
+==============
 
-Peripherals are represented as child nodes of the DSI host's node. Properties
-described here apply to all DSI peripherals, but individual bindings may want
-to define additional, device-specific properties.
+Peripherals with DSI as control bus, or no control bus
+------------------------------------------------------
+
+Peripherals with the DSI bus as the primary control bus, or peripherals with
+no control bus but use the DSI bus to transmit pixel data are represented
+as child nodes of the DSI host's node. Properties described here apply to all
+DSI peripherals, but individual bindings may want to define additional,
+device-specific properties.
 
 Required properties:
 - reg: The virtual channel number of a DSI peripheral. Must be in the range

@@ -49,9 +54,25 @@ case two alternative representations can be chosen:
   property is the number of the first virtual channel and the second cell is
   the number of consecutive virtual channels.
 
-Example
--------
+Peripherals with a different control bus
+----------------------------------------
+
+There are peripherals that have I2C/SPI (or some other non-DSI bus) as the
+primary control bus, but are also connected to a DSI bus (mostly for the data
+path). Connections between such peripherals and a DSI host can be represented

Documentation/devicetree/bindings/display/mipi-dsi-bus.txt +39 -0

+  a DSI host using of-graph bindings.
+
+1)
 	dsi-host {
...

@@ -67,6 +88,7 @@ Example
...
 	};
 
+2)
 	dsi-host {
...

@@ -82,6 +104,7 @@ Example
...
 	};
 
+3)
 	dsi-host {
...

@@ -96,3 +119,37 @@ Example
 
...
 	};
+
+4)
+	i2c-host {
+		...
+
+		dsi-bridge@35 {
+			compatible = "...";
+			reg = <0x35>;
+
+			ports {
+				...
+
+				port {
+					bridge_mipi_in: endpoint {
+						remote-endpoint = <&host_mipi_out>;
+					};
+				};
+			};
+		};
+	};
+
+	dsi-host {
+		...
+
+		ports {
+			...
+
+			port {
+				host_mipi_out: endpoint {
+					remote-endpoint = <&bridge_mipi_in>;
+				};
+			};
+		};
+	};

Signature · 2 lines

The Qualcomm Innovation Center, Inc. is a member of the Code Aurora Forum,
hosted by The Linux Foundation

Add binding info for peripherals that support dual-channel DSI. Add corresponding optional bindings for DSI host controllers that may be configured in this mode. Add an example of an I2C controlled device operating in dual-channel DSI mode.

Signed-off-by Archit Taneja <[email protected]>

---
v2:
- Specify that clock-master is a boolean property.
- Drop/add unit-address and #*-cells where applicable.

  .../devicetree/bindings/display/mipi-dsi-bus.txt   | 80 ++++++++++++++++++++++
  1 file changed, 80 insertions(+)

Documentation/devicetree/bindings/display/mipi-dsi-bus.txt +70 -0

index 94fb72cb916f..7a3abbedb3fa 100644
@@ -29,6 +29,13 @@ Required properties:
  - #size-cells: Should be 0. There are cases where it makes sense to use a
    different value here. See below.
 
+Optional properties:
+- clock-master: boolean. Should be enabled if the host is being used in
+  conjunction with another DSI host to drive the same peripheral. Hardware
+  supporting such a configuration generally requires the data on both the busses
+  to be driven by the same clock. Only the DSI host instance controlling this
+  clock should contain this property.
+
  DSI peripheral
  ==============
 

@@ -62,6 +69,16 @@ primary control bus, but are also connected to a DSI bus (mostly for the data
  path). Connections between such peripherals and a DSI host can be represented
  using the graph bindings [1], [2].
 
+Peripherals that support dual channel DSI
+-----------------------------------------
+
+Peripherals with higher bandwidth requirements can be connected to 2 DSI
+busses. Each DSI bus/channel drives some portion of the pixel data (generally
+left/right half of each line of the display, or even/odd lines of the display).
+The graph bindings should be used to represent the multiple DSI busses that are
+connected to this peripheral. Each DSI host's output endpoint can be linked to
+an input endpoint of the DSI peripheral.
+
  [1] Documentation/devicetree/bindings/graph.txt
  [2] Documentation/devicetree/bindings/media/video-interfaces.txt
 

@@ -71,6 +88,8 @@ Examples
    with different virtual channel configurations.
  - (4) is an example of a peripheral on a I2C control bus connected with to
    a DSI host using of-graph bindings.
+- (5) is an example of 2 DSI hosts driving a dual-channel DSI peripheral,
+  which uses I2C as its primary control bus.
 
  1)
  	dsi-host {

@@ -153,3 +172,64 @@ Examples
  			};
  		};
  	};
+
+5)
+	i2c-host {
+		dsi-bridge@35 {
+			compatible = "...";
+			reg = <0x35>;
+
+			ports {
+				#address-cells = <1>;
+				#size-cells = <0>;
+
+				port@0 {
+					reg = <0>;
+					dsi0_in: endpoint {
+						remote-endpoint = <&dsi0_out>;
+					};
+				};
+
+				port@1 {
+					reg = <1>;
+					dsi1_in: endpoint {
+						remote-endpoint = <&dsi1_out>;
+					};
+				};
+			};
+		};
+	};
+
+	dsi0-host {
+		...
+
+		/*
+		 * this DSI instance drives the clock for both the host
+		 * controllers
+		 */
+		clock-master;
+
+		ports {
+			...
+
+			port {
+				dsi0_out: endpoint {
+					remote-endpoint = <&dsi0_in>;
+				};
+			};
+		};
+	};
+
+	dsi1-host {
+		...
+

Add binding info for peripherals that support dual-channel DSI. Add corresponding optional bindings for DSI host controllers that may be configured in this mode. Add an example of an I2C controlled device operating in dual-channel DSI mode.

Signed-off-by Archit Taneja <[email protected]>

---
v2:
- Specify that clock-master is a boolean property.
- Drop/add unit-address and #*-cells where applicable.

 .../devicetree/bindings/display/mipi-dsi-bus.txt   | 80 ++++++++++++++++++++++
 1 file changed, 80 insertions(+)

Documentation/devicetree/bindings/display/mipi-dsi-bus.txt +80 -0

index 94fb72cb916f..7a3abbedb3fa 100644
@@ -29,6 +29,13 @@ Required properties:
 - #size-cells: Should be 0. There are cases where it makes sense to use a
   different value here. See below.
 
+Optional properties:
+- clock-master: boolean. Should be enabled if the host is being used in
+  conjunction with another DSI host to drive the same peripheral. Hardware
+  supporting such a configuration generally requires the data on both the busses
+  to be driven by the same clock. Only the DSI host instance controlling this
+  clock should contain this property.
+
 DSI peripheral
 ==============
 

@@ -62,6 +69,16 @@ primary control bus, but are also connected to a DSI bus (mostly for the data
 path). Connections between such peripherals and a DSI host can be represented
 using the graph bindings [1], [2].
 
+Peripherals that support dual channel DSI
+-----------------------------------------
+
+Peripherals with higher bandwidth requirements can be connected to 2 DSI
+busses. Each DSI bus/channel drives some portion of the pixel data (generally
+left/right half of each line of the display, or even/odd lines of the display).
+The graph bindings should be used to represent the multiple DSI busses that are
+connected to this peripheral. Each DSI host's output endpoint can be linked to
+an input endpoint of the DSI peripheral.
+
 [1] Documentation/devicetree/bindings/graph.txt
 [2] Documentation/devicetree/bindings/media/video-interfaces.txt
 

@@ -71,6 +88,8 @@ Examples
   with different virtual channel configurations.
 - (4) is an example of a peripheral on a I2C control bus connected with to
   a DSI host using of-graph bindings.
+- (5) is an example of 2 DSI hosts driving a dual-channel DSI peripheral,
+  which uses I2C as its primary control bus.
 
 1)
 	dsi-host {

@@ -153,3 +172,64 @@ Examples
 			};
 		};
 	};
+
+5)
+	i2c-host {
+		dsi-bridge@35 {
+			compatible = "...";
+			reg = <0x35>;
+
+			ports {
+				#address-cells = <1>;
+				#size-cells = <0>;
+
+				port@0 {
+					reg = <0>;
+					dsi0_in: endpoint {
+						remote-endpoint = <&dsi0_out>;
+					};
+				};
+
+				port@1 {
+					reg = <1>;
+					dsi1_in: endpoint {
+						remote-endpoint = <&dsi1_out>;
+					};
+				};
+			};
+		};
+	};
+
+	dsi0-host {
+		...
+
+		/*
+		 * this DSI instance drives the clock for both the host
+		 * controllers
+		 */
+		clock-master;
+
+		ports {
+			...
+
+			port {
+				dsi0_out: endpoint {
+					remote-endpoint = <&dsi0_in>;
+				};
+			};
+		};
+	};
+
+	dsi1-host {
+		...
+
+		ports {
+			...
+
+			port {
+				dsi1_out: endpoint {
+					remote-endpoint = <&dsi1_in>;
+				};
+			};
+		};
+	};

Signature · 2 lines

The Qualcomm Innovation Center, Inc. is a member of the Code Aurora Forum,
hosted by The Linux Foundation

Add binding info for peripherals that support dual-channel DSI. Add corresponding optional bindings for DSI host controllers that may be configured in this mode. Add an example of an I2C controlled device operating in dual-channel DSI mode.

Signed-off-by Archit Taneja <[email protected]>

---
v2:
- Specify that clock-master is a boolean property.
- Drop/add unit-address and #*-cells where applicable.

 .../devicetree/bindings/display/mipi-dsi-bus.txt   | 80 ++++++++++++++++++++++
 1 file changed, 80 insertions(+)

Documentation/devicetree/bindings/display/mipi-dsi-bus.txt +70 -0

index 94fb72cb916f..7a3abbedb3fa 100644
@@ -29,6 +29,13 @@ Required properties:
 - #size-cells: Should be 0. There are cases where it makes sense to use a
   different value here. See below.
 
+Optional properties:
+- clock-master: boolean. Should be enabled if the host is being used in
+  conjunction with another DSI host to drive the same peripheral. Hardware
+  supporting such a configuration generally requires the data on both the busses
+  to be driven by the same clock. Only the DSI host instance controlling this
+  clock should contain this property.
+
 DSI peripheral
 ==============
 

@@ -62,6 +69,16 @@ primary control bus, but are also connected to a DSI bus (mostly for the data
 path). Connections between such peripherals and a DSI host can be represented
 using the graph bindings [1], [2].
 
+Peripherals that support dual channel DSI
+-----------------------------------------
+
+Peripherals with higher bandwidth requirements can be connected to 2 DSI
+busses. Each DSI bus/channel drives some portion of the pixel data (generally
+left/right half of each line of the display, or even/odd lines of the display).
+The graph bindings should be used to represent the multiple DSI busses that are
+connected to this peripheral. Each DSI host's output endpoint can be linked to
+an input endpoint of the DSI peripheral.
+
 [1] Documentation/devicetree/bindings/graph.txt
 [2] Documentation/devicetree/bindings/media/video-interfaces.txt
 

@@ -71,6 +88,8 @@ Examples
   with different virtual channel configurations.
 - (4) is an example of a peripheral on a I2C control bus connected with to
   a DSI host using of-graph bindings.
+- (5) is an example of 2 DSI hosts driving a dual-channel DSI peripheral,
+  which uses I2C as its primary control bus.
 
 1)
 	dsi-host {

@@ -153,3 +172,64 @@ Examples
 			};
 		};
 	};
+
+5)
+	i2c-host {
+		dsi-bridge@35 {
+			compatible = "...";
+			reg = <0x35>;
+
+			ports {
+				#address-cells = <1>;
+				#size-cells = <0>;
+
+				port@0 {
+					reg = <0>;
+					dsi0_in: endpoint {
+						remote-endpoint = <&dsi0_out>;
+					};
+				};
+
+				port@1 {
+					reg = <1>;
+					dsi1_in: endpoint {
+						remote-endpoint = <&dsi1_out>;
+					};
+				};
+			};
+		};
+	};
+
+	dsi0-host {
+		...
+
+		/*
+		 * this DSI instance drives the clock for both the host
+		 * controllers
+		 */
+		clock-master;
+
+		ports {
+			...
+
+			port {
+				dsi0_out: endpoint {
+					remote-endpoint = <&dsi0_in>;
+				};
+			};
+		};
+	};
+
+	dsi1-host {
+		...
+

Am Donnerstag, 18. Januar 2018, 05:53:55 CEST schrieb Archit Taneja:

Quoted 6 lines

Add binding info for peripherals that support dual-channel DSI. Add corresponding optional bindings for DSI host controllers that may be configured in this mode. Add an example of an I2C controlled device operating in dual-channel DSI mode.

Signed-off-by Archit Taneja <[email protected]>

Looks like a great solution for that problem, so

Reviewed-by Heiko Stuebner <[email protected]>

As I'm looking into that for my rk3399-scarlet device right now and couldn't find this patchset in the kernel yet, is it planned to merge or refresh these binding changes or were problems encountered.

At least an Ack/Review from Rob seems to be missing.

Heiko

Quoted 108 lines

---
v2:
- Specify that clock-master is a boolean property.
- Drop/add unit-address and #*-cells where applicable.

  .../devicetree/bindings/display/mipi-dsi-bus.txt   | 80 ++++++++++++++++++++++
  1 file changed, 80 insertions(+)

Documentation/devicetree/bindings/display/mipi-dsi-bus.txt +70 -0

index 94fb72cb916f..7a3abbedb3fa 100644
@@ -29,6 +29,13 @@ Required properties:
  - #size-cells: Should be 0. There are cases where it makes sense to use a
    different value here. See below.
 
+Optional properties:
+- clock-master: boolean. Should be enabled if the host is being used in
+  conjunction with another DSI host to drive the same peripheral. Hardware
+  supporting such a configuration generally requires the data on both the busses
+  to be driven by the same clock. Only the DSI host instance controlling this
+  clock should contain this property.
+
  DSI peripheral
  ==============
 

@@ -62,6 +69,16 @@ primary control bus, but are also connected to a DSI bus (mostly for the data
  path). Connections between such peripherals and a DSI host can be represented
  using the graph bindings [1], [2].
 
+Peripherals that support dual channel DSI
+-----------------------------------------
+
+Peripherals with higher bandwidth requirements can be connected to 2 DSI
+busses. Each DSI bus/channel drives some portion of the pixel data (generally
+left/right half of each line of the display, or even/odd lines of the display).
+The graph bindings should be used to represent the multiple DSI busses that are
+connected to this peripheral. Each DSI host's output endpoint can be linked to
+an input endpoint of the DSI peripheral.
+
  [1] Documentation/devicetree/bindings/graph.txt
  [2] Documentation/devicetree/bindings/media/video-interfaces.txt
 

@@ -71,6 +88,8 @@ Examples
    with different virtual channel configurations.
  - (4) is an example of a peripheral on a I2C control bus connected with to
    a DSI host using of-graph bindings.
+- (5) is an example of 2 DSI hosts driving a dual-channel DSI peripheral,
+  which uses I2C as its primary control bus.
 
  1)
  	dsi-host {

@@ -153,3 +172,64 @@ Examples
  			};
  		};
  	};
+
+5)
+	i2c-host {
+		dsi-bridge@35 {
+			compatible = "...";
+			reg = <0x35>;
+
+			ports {
+				#address-cells = <1>;
+				#size-cells = <0>;
+
+				port@0 {
+					reg = <0>;
+					dsi0_in: endpoint {
+						remote-endpoint = <&dsi0_out>;
+					};
+				};
+
+				port@1 {
+					reg = <1>;
+					dsi1_in: endpoint {
+						remote-endpoint = <&dsi1_out>;
+					};
+				};
+			};
+		};
+	};
+
+	dsi0-host {
+		...
+
+		/*
+		 * this DSI instance drives the clock for both the host
+		 * controllers
+		 */
+		clock-master;
+
+		ports {
+			...
+
+			port {
+				dsi0_out: endpoint {
+					remote-endpoint = <&dsi0_in>;
+				};
+			};
+		};
+	};
+
+	dsi1-host {
+		...
+

Documentation/devicetree/bindings/display/mipi-dsi-bus.txt Examples · line 235

Quoted 10 lines

Documentation/devicetree/bindings/display/mipi-dsi-bus.txt +10 -0

+		ports {
+			...
+
+			port {
+				dsi1_out: endpoint {
+					remote-endpoint = <&dsi1_in>;
+				};
+			};
+		};
+	};

--
To unsubscribe from this list: send the line "unsubscribe linux-arm-msm" in the body of a message to [email protected]
More majordomo info at http://vger.kernel.org/majordomo-info.html

Quoted 15 lines

Am Donnerstag, 18. Januar 2018, 05:53:55 CEST schrieb Archit Taneja:

Quoted 6 lines

Add binding info for peripherals that support dual-channel DSI. Add corresponding optional bindings for DSI host controllers that may be configured in this mode. Add an example of an I2C controlled device operating in dual-channel DSI mode.

Signed-off-by Archit Taneja <[email protected]>

Looks like a great solution for that problem, so

Reviewed-by Heiko Stuebner <[email protected]>

As I'm looking into that for my rk3399-scarlet device right now and couldn't find this patchset in the kernel yet, is it planned to merge or refresh these binding changes or were problems encountered.

Quoted 1 line

At least an Ack/Review from Rob seems to be missing.

I forgot about these patches. Rob had reviewed the first one in the set the second one still needed an Ack. I'll post a v3 that adds the Reviewed-bys and fixes a small typo.

Quoted 136 lines

Quoted 136 lines

---
v2:
- Specify that clock-master is a boolean property.
- Drop/add unit-address and #*-cells where applicable.

  .../devicetree/bindings/display/mipi-dsi-bus.txt   | 80

++++++++++++++++++++++ 1 file changed, 80 insertions(+)

(quoted diff) +0 -0

b/Documentation/devicetree/bindings/display/mipi-dsi-bus.txt index 94fb72cb916f..7a3abbedb3fa 100644

Documentation/devicetree/bindings/display/mipi-dsi-bus.txt +3 -0

@@ -29,6 +29,13 @@ Required properties:
  - #size-cells: Should be 0. There are cases where it makes sense to use
  a
 
    different value here. See below.
 
+Optional properties:
+- clock-master: boolean. Should be enabled if the host is being used in
+  conjunction with another DSI host to drive the same peripheral.

Hardware
+ supporting such a configuration generally requires the data on both the busses + to be driven by the same clock. Only the DSI host instance controlling this + clock should contain this property.

Documentation/devicetree/bindings/display/mipi-dsi-bus.txt +1 -0

+
 
  DSI peripheral
  ==============

@@ -62,6 +69,16 @@ primary control bus, but are also connected to a DSI

bus (mostly for the data>>
path). Connections between such peripherals and a DSI host can be represented using the graph bindings [1], [2].

Documentation/devicetree/bindings/display/mipi-dsi-bus.txt +5 -0

+Peripherals that support dual channel DSI
+-----------------------------------------
+
+Peripherals with higher bandwidth requirements can be connected to 2 DSI
+busses. Each DSI bus/channel drives some portion of the pixel data

(generally +left/right half of each line of the display, or even/odd lines of the display). +The graph bindings should be used to represent the multiple DSI busses that are +connected to this peripheral. Each DSI host's output endpoint can be linked to +an input endpoint of the DSI peripheral.

Documentation/devicetree/bindings/display/mipi-dsi-bus.txt +2 -0

+
 
  [1] Documentation/devicetree/bindings/graph.txt
  [2] Documentation/devicetree/bindings/media/video-interfaces.txt

@@ -71,6 +88,8 @@ Examples
 
    with different virtual channel configurations.
 
  - (4) is an example of a peripheral on a I2C control bus connected with
  to
 
    a DSI host using of-graph bindings.
 
+- (5) is an example of 2 DSI hosts driving a dual-channel DSI

peripheral,

Documentation/devicetree/bindings/display/mipi-dsi-bus.txt +60 -0

+  which uses I2C as its primary control bus.
 
  1)
 
  	dsi-host {

@@ -153,3 +172,64 @@ Examples
 
  			};
 
  		};
 
  	};
 
+
+5)
+	i2c-host {
+		dsi-bridge@35 {
+			compatible = "...";
+			reg = <0x35>;
+
+			ports {
+				#address-cells = <1>;
+				#size-cells = <0>;
+
+				port@0 {
+					reg = <0>;
+					dsi0_in: endpoint {
+						remote-endpoint = <&dsi0_out>;
+					};
+				};
+
+				port@1 {
+					reg = <1>;
+					dsi1_in: endpoint {
+						remote-endpoint = <&dsi1_out>;
+					};
+				};
+			};
+		};
+	};
+
+	dsi0-host {
+		...
+
+		/*
+		 * this DSI instance drives the clock for both the host
+		 * controllers
+		 */
+		clock-master;
+
+		ports {
+			...
+
+			port {
+				dsi0_out: endpoint {
+					remote-endpoint = <&dsi0_in>;
+				};
+			};
+		};
+	};
+
+	dsi1-host {
+		...
+
+		ports {
+			...
+
+			port {
+				dsi1_out: endpoint {
+					remote-endpoint = <&dsi1_in>;
+				};
+			};

Quoted 8 lines

Documentation/devicetree/bindings/display/mipi-dsi-bus.txt

Quoted 2 lines

Documentation/devicetree/bindings/display/mipi-dsi-bus.txt +2 -0

+		};
+	};

--
To unsubscribe from this list: send the line "unsubscribe linux-arm-msm" in
the body of a message to [email protected]
More majordomo info at http://vger.kernel.org/majordomo-info.html

Am Mittwoch, 6. Juni 2018, 07:59:29 CEST schrieb Archit Taneja:

Quoted 16 lines · Heiko Stuebner

Quoted 15 lines

Am Donnerstag, 18. Januar 2018, 05:53:55 CEST schrieb Archit Taneja:

Quoted 6 lines

Add binding info for peripherals that support dual-channel DSI. Add corresponding optional bindings for DSI host controllers that may be configured in this mode. Add an example of an I2C controlled device operating in dual-channel DSI mode.

Signed-off-by Archit Taneja <[email protected]>

Looks like a great solution for that problem, so

Reviewed-by Heiko Stuebner <[email protected]>

As I'm looking into that for my rk3399-scarlet device right now and couldn't find this patchset in the kernel yet, is it planned to merge or refresh these binding changes or were problems encountered.

Quoted 5 lines

Quoted 1 line

At least an Ack/Review from Rob seems to be missing.

I forgot about these patches. Rob had reviewed the first one in the set the second one still needed an Ack. I'll post a v3 that adds the Reviewed-bys and fixes a small typo.

very nice ... because it looks like yesterday I managed to make the Rockchip dsi work in dual mode following this.

But one question came up, do you really want two input ports on the panel side? I.e. hardware-wise, I guess the panel will have one 8-lane or so input thatonly gets split up on the soc side onto 2 dsi controllers?

So right now I'm operating with a devicetree like

&mipi_dsi {
         status = "okay";
         clock-master;

         ports {
                 mipi_out: port@1 {
                         reg = <1>;

                         mipi_out_panel: endpoint {
                                 remote-endpoint = <&mipi_in_panel>;
                         };
                 };
         };

         mipi_panel: panel@0 {
		  compatible = "innolux,p097pfg";
                 reg = <0>;
                 backlight = <&backlight>;
                 enable-gpios = <&gpio4 25 GPIO_ACTIVE_HIGH>;
                 pinctrl-names = "default";
                 pinctrl-0 = <&display_rst_l>;

                 port {
                         #address-cells = <1>;
                         #size-cells = <0>;

                         mipi_in_panel: endpoint@0 {
                                 reg = <0>;
                                 remote-endpoint = <&mipi_out_panel>;
                         };

                         mipi1_in_panel: endpoint@1 {
                                 reg = <1>;
                                 remote-endpoint = <&mipi1_out_panel>;
                         };
                 };
         };
};

&mipi_dsi1 {
         status = "okay";

         ports {
                 mipi1_out: port@1 {
                         reg = <1>;

                         mipi1_out_panel: endpoint {
                                 remote-endpoint = <&mipi1_in_panel>;
                         };
                 };
         };
};

I guess it is a matter of preference on what reflects the hardware best, so maybe that's Robs call?

Heiko

Quoted 144 lines

Quoted 144 lines

Quoted 128 lines

---
v2:
- Specify that clock-master is a boolean property.
- Drop/add unit-address and #*-cells where applicable.

   .../devicetree/bindings/display/mipi-dsi-bus.txt   | 80

++++++++++++++++++++++ 1 file changed, 80 insertions(+)

(quoted diff) +0 -0

b/Documentation/devicetree/bindings/display/mipi-dsi-bus.txt index 94fb72cb916f..7a3abbedb3fa 100644

Documentation/devicetree/bindings/display/mipi-dsi-bus.txt +3 -0

@@ -29,6 +29,13 @@ Required properties:
   - #size-cells: Should be 0. There are cases where it makes sense to use
   a
 
     different value here. See below.
 
+Optional properties:
+- clock-master: boolean. Should be enabled if the host is being used in
+  conjunction with another DSI host to drive the same peripheral.

Hardware
+ supporting such a configuration generally requires the data on both the busses + to be driven by the same clock. Only the DSI host instance controlling this + clock should contain this property.

Documentation/devicetree/bindings/display/mipi-dsi-bus.txt +1 -0

+
 
   DSI peripheral
   ==============

@@ -62,6 +69,16 @@ primary control bus, but are also connected to a DSI

bus (mostly for the data>>
path). Connections between such peripherals and a DSI host can be represented using the graph bindings [1], [2].

Documentation/devicetree/bindings/display/mipi-dsi-bus.txt +5 -0

+Peripherals that support dual channel DSI
+-----------------------------------------
+
+Peripherals with higher bandwidth requirements can be connected to 2 DSI
+busses. Each DSI bus/channel drives some portion of the pixel data

(generally +left/right half of each line of the display, or even/odd lines of the display). +The graph bindings should be used to represent the multiple DSI busses that are +connected to this peripheral. Each DSI host's output endpoint can be linked to +an input endpoint of the DSI peripheral.

Documentation/devicetree/bindings/display/mipi-dsi-bus.txt +2 -0

+
 
   [1] Documentation/devicetree/bindings/graph.txt
   [2] Documentation/devicetree/bindings/media/video-interfaces.txt

@@ -71,6 +88,8 @@ Examples
 
     with different virtual channel configurations.
 
   - (4) is an example of a peripheral on a I2C control bus connected with
   to
 
     a DSI host using of-graph bindings.
 
+- (5) is an example of 2 DSI hosts driving a dual-channel DSI

peripheral,

Documentation/devicetree/bindings/display/mipi-dsi-bus.txt +52 -0

+  which uses I2C as its primary control bus.
 
   1)
 
   	dsi-host {

@@ -153,3 +172,64 @@ Examples
 
   			};
 
   		};
 
   	};
 
+
+5)
+	i2c-host {
+		dsi-bridge@35 {
+			compatible = "...";
+			reg = <0x35>;
+
+			ports {
+				#address-cells = <1>;
+				#size-cells = <0>;
+
+				port@0 {
+					reg = <0>;
+					dsi0_in: endpoint {
+						remote-endpoint = <&dsi0_out>;
+					};
+				};
+
+				port@1 {
+					reg = <1>;
+					dsi1_in: endpoint {
+						remote-endpoint = <&dsi1_out>;
+					};
+				};
+			};
+		};
+	};
+
+	dsi0-host {
+		...
+
+		/*
+		 * this DSI instance drives the clock for both the host
+		 * controllers
+		 */
+		clock-master;
+
+		ports {
+			...
+
+			port {
+				dsi0_out: endpoint {
+					remote-endpoint = <&dsi0_in>;
+				};
+			};
+		};
+	};
+
+	dsi1-host {
+		...
+

Documentation/devicetree/bindings/display/mipi-dsi-bus.txt Examples · line 235

Quoted 10 lines

Documentation/devicetree/bindings/display/mipi-dsi-bus.txt +10 -0

+		ports {
+			...
+
+			port {
+				dsi1_out: endpoint {
+					remote-endpoint = <&dsi1_in>;
+				};
+			};
+		};
+	};

--
To unsubscribe from this list: send the line "unsubscribe linux-arm-msm" in
the body of a message to [email protected]
More majordomo info at http://vger.kernel.org/majordomo-info.html

Quoted 23 lines

Am Mittwoch, 6. Juni 2018, 07:59:29 CEST schrieb Archit Taneja:

Quoted 16 lines · Heiko Stuebner

Quoted 15 lines

Am Donnerstag, 18. Januar 2018, 05:53:55 CEST schrieb Archit Taneja:

Quoted 6 lines

Add binding info for peripherals that support dual-channel DSI. Add corresponding optional bindings for DSI host controllers that may be configured in this mode. Add an example of an I2C controlled device operating in dual-channel DSI mode.

Signed-off-by Archit Taneja <[email protected]>

Looks like a great solution for that problem, so

Reviewed-by Heiko Stuebner <[email protected]>

As I'm looking into that for my rk3399-scarlet device right now and couldn't find this patchset in the kernel yet, is it planned to merge or refresh these binding changes or were problems encountered.

Quoted 5 lines

Quoted 1 line

At least an Ack/Review from Rob seems to be missing.

I forgot about these patches. Rob had reviewed the first one in the set the second one still needed an Ack. I'll post a v3 that adds the Reviewed-bys and fixes a small typo.

Quoted 6 lines

very nice ... because it looks like yesterday I managed to make the Rockchip dsi work in dual mode following this.

But one question came up, do you really want two input ports on the panel side? I.e. hardware-wise, I guess the panel will have one 8-lane or so input thatonly gets split up on the soc side onto 2 dsi controllers?

I think all dual DSI panels actually have 2 DSI controllers/parsers within them, one on each port. The MIPI DSI spec doesn't support 8 lanes. Also, the pixels coming out of the host are distributed among the lanes differently than what would have been the case with a 'theoretical' 8 lane receiver.

Other than that, some dual DSI panels only accept DSI commands on the 'master' port, where as others expect the same command to be sent across both the ports.

Therefore, I think it's better to represent dual DSI panels having 2 DSI input ports.

Your DT looks good to me.

Quoted 230 lines

So right now I'm operating with a devicetree like

&mipi_dsi {

         status = "okay";
         clock-master;
         
         ports {
         
                 mipi_out: port@1 {
                 
                         reg = <1>;
                         
                         mipi_out_panel: endpoint {
                         
                                 remote-endpoint = <&mipi_in_panel>;
                         
                         };
                 
                 };
         
         };
         
         mipi_panel: panel@0 {
		  
		  compatible = "innolux,p097pfg";
		  
                 reg = <0>;
                 backlight = <&backlight>;
                 enable-gpios = <&gpio4 25 GPIO_ACTIVE_HIGH>;
                 pinctrl-names = "default";
                 pinctrl-0 = <&display_rst_l>;
                 
                 port {
                 
                         #address-cells = <1>;
                         #size-cells = <0>;
                         
                         mipi_in_panel: endpoint@0 {
                         
                                 reg = <0>;
                                 remote-endpoint = <&mipi_out_panel>;
                         
                         };
                         
                         mipi1_in_panel: endpoint@1 {
                         
                                 reg = <1>;
                                 remote-endpoint = <&mipi1_out_panel>;
                         
                         };
                 
                 };
         
         };

};

&mipi_dsi1 {

         status = "okay";
         
         ports {
         
                 mipi1_out: port@1 {
                 
                         reg = <1>;
                         
                         mipi1_out_panel: endpoint {
                         
                                 remote-endpoint = <&mipi1_in_panel>;
                         
                         };
                 
                 };
         
         };

};

I guess it is a matter of preference on what reflects the hardware best, so maybe that's Robs call?

Heiko

Quoted 143 lines

Quoted 143 lines

Quoted 143 lines

---
v2:
- Specify that clock-master is a boolean property.
- Drop/add unit-address and #*-cells where applicable.

   .../devicetree/bindings/display/mipi-dsi-bus.txt   | 80

++++++++++++++++++++++ 1 file changed, 80 insertions(+)

(quoted diff) +0 -0

b/Documentation/devicetree/bindings/display/mipi-dsi-bus.txt index 94fb72cb916f..7a3abbedb3fa 100644

Documentation/devicetree/bindings/display/mipi-dsi-bus.txt +2 -0

@@ -29,6 +29,13 @@ Required properties:
   - #size-cells: Should be 0. There are cases where it makes sense to
   use
   a
 
     different value here. See below.
 
+Optional properties:
+- clock-master: boolean. Should be enabled if the host is being used

in
+ conjunction with another DSI host to drive the same peripheral. Hardware
+ supporting such a configuration generally requires the data on both the busses + to be driven by the same clock. Only the DSI host instance
controlling this + clock should contain this property.

Documentation/devicetree/bindings/display/mipi-dsi-bus.txt +1 -0

+
 
   DSI peripheral
   ==============

@@ -62,6 +69,16 @@ primary control bus, but are also connected to a DSI

bus (mostly for the data>>

path). Connections between such peripherals and a DSI host can be represented using the graph bindings [1], [2].

Documentation/devicetree/bindings/display/mipi-dsi-bus.txt +4 -0

+Peripherals that support dual channel DSI
+-----------------------------------------
+
+Peripherals with higher bandwidth requirements can be connected to 2

DSI
+busses. Each DSI bus/channel drives some portion of the pixel data (generally +left/right half of each line of the display, or even/odd lines of the display). +The graph bindings should be used to represent the multiple DSI busses that are +connected to this peripheral. Each DSI
host's output endpoint can be linked to +an input endpoint of the DSI peripheral.

Documentation/devicetree/bindings/display/mipi-dsi-bus.txt +2 -0

+
 
   [1] Documentation/devicetree/bindings/graph.txt
   [2] Documentation/devicetree/bindings/media/video-interfaces.txt

@@ -71,6 +88,8 @@ Examples
 
     with different virtual channel configurations.
 
   - (4) is an example of a peripheral on a I2C control bus connected
   with
   to
 
     a DSI host using of-graph bindings.
 
+- (5) is an example of 2 DSI hosts driving a dual-channel DSI

peripheral,

Documentation/devicetree/bindings/display/mipi-dsi-bus.txt +60 -0

+  which uses I2C as its primary control bus.
 
   1)
 
   	dsi-host {

@@ -153,3 +172,64 @@ Examples
 
   			};
 
   		};
 
   	};
 
+
+5)
+	i2c-host {
+		dsi-bridge@35 {
+			compatible = "...";
+			reg = <0x35>;
+
+			ports {
+				#address-cells = <1>;
+				#size-cells = <0>;
+
+				port@0 {
+					reg = <0>;
+					dsi0_in: endpoint {
+						remote-endpoint = <&dsi0_out>;
+					};
+				};
+
+				port@1 {
+					reg = <1>;
+					dsi1_in: endpoint {
+						remote-endpoint = <&dsi1_out>;
+					};
+				};
+			};
+		};
+	};
+
+	dsi0-host {
+		...
+
+		/*
+		 * this DSI instance drives the clock for both the host
+		 * controllers
+		 */
+		clock-master;
+
+		ports {
+			...
+
+			port {
+				dsi0_out: endpoint {
+					remote-endpoint = <&dsi0_in>;
+				};
+			};
+		};
+	};
+
+	dsi1-host {
+		...
+
+		ports {
+			...
+
+			port {
+				dsi1_out: endpoint {
+					remote-endpoint = <&dsi1_in>;
+				};
+			};

Quoted 8 lines

Quoted 8 lines

Quoted 8 lines

Documentation/devicetree/bindings/display/mipi-dsi-bus.txt

Quoted 2 lines

Documentation/devicetree/bindings/display/mipi-dsi-bus.txt +2 -0

+		};
+	};

--
To unsubscribe from this list: send the line "unsubscribe linux-arm-msm" in
the body of a message to [email protected]
More majordomo info at http://vger.kernel.org/majordomo-info.html

Hi Archit,

Am Mittwoch, 6. Juni 2018, 12:21:16 CEST schrieb Archit Taneja:

Quoted 41 lines · Heiko Stübner

Quoted 23 lines

Am Mittwoch, 6. Juni 2018, 07:59:29 CEST schrieb Archit Taneja:

Quoted 16 lines · Heiko Stuebner

Quoted 15 lines

Am Donnerstag, 18. Januar 2018, 05:53:55 CEST schrieb Archit Taneja:

Quoted 6 lines

Add binding info for peripherals that support dual-channel DSI. Add corresponding optional bindings for DSI host controllers that may be configured in this mode. Add an example of an I2C controlled device operating in dual-channel DSI mode.

Signed-off-by Archit Taneja <[email protected]>

Looks like a great solution for that problem, so

Reviewed-by Heiko Stuebner <[email protected]>

As I'm looking into that for my rk3399-scarlet device right now and couldn't find this patchset in the kernel yet, is it planned to merge or refresh these binding changes or were problems encountered.

Quoted 5 lines

Quoted 1 line

At least an Ack/Review from Rob seems to be missing.

I forgot about these patches. Rob had reviewed the first one in the set the second one still needed an Ack. I'll post a v3 that adds the Reviewed-bys and fixes a small typo.

Quoted 6 lines

very nice ... because it looks like yesterday I managed to make the Rockchip dsi work in dual mode following this.

But one question came up, do you really want two input ports on the panel side? I.e. hardware-wise, I guess the panel will have one 8-lane or so input thatonly gets split up on the soc side onto 2 dsi controllers?

I think all dual DSI panels actually have 2 DSI controllers/parsers within them, one on each port. The MIPI DSI spec doesn't support 8 lanes. Also, the pixels coming out of the host are distributed among the lanes differently than what would have been the case with a 'theoretical' 8 lane receiver.

Other than that, some dual DSI panels only accept DSI commands on the 'master' port, where as others expect the same command to be sent across both the ports.

Quoted 4 lines

Therefore, I think it's better to represent dual DSI panels having 2 DSI input ports.

Your DT looks good to me.

Hmm, that doesn't match up then ;-) ... as my dt uses 2 endpoints in one port for the dsi-links.

The issue I see with using ports and not endpoints for dual-dsi links is with distinguishing between input and output ports.

For a panel that's easy, as you every port will be an input port and if you have 2, it's supposed dual-dsi. But for example I guess there might exist some dual-dsi-to-something bridges, where you would end up with say 3 (or even more) ports ... two dual-dsi inputs and 1 or more outputs.

While the following argument might not be 100% valid from a dt-purity standpoint implementing this might get hairy for _any_ operating system, as you will need each panel/bridge to tell what the ports are used for.

I.e. in my endpoint based mapping, right now I have this nice and generic WIP function to parse the of_graph and get the master+slave nodes:

https://github.com/mmind/linux-rockchip/blob/tmp/rk3399-gru-bob-scarlet/drivers/gpu/drm/rockchip/dw-mipi-dsi-rockchip.c#L697
[0]

So I guess my proposal would be to have one port for inputs and one port for outputs for dsi peripherals, with possibly multiple endpoints in each.

Heiko

[0] github seems to have reliability problems, so for reference my parsing function:

static int dw_mipi_dsi_is_dual(struct dw_mipi_dsi_rockchip *dsi,
struct device_node **master, struct device_node **slave)

{
	struct device_node *local_ep, *remote_port, *ep;
	struct device_node *ctrls[2] = { NULL, NULL };
	int num = 0, ret = 0, idx;

	/* get local panel endpoint of the dsi controller */
	local_ep = of_graph_get_endpoint_by_regs(dsi->dev->of_node, 1, 0);
	if (!local_ep) {
		DRM_DEV_ERROR(dsi->dev, "couldn't find local panel endpoint\n");
		return -ENXIO;
	}

	/* get panel port */
	remote_port = of_graph_get_remote_port_parent(local_ep);
	of_node_put(local_ep);
	if (!remote_port) {
		DRM_DEV_ERROR(dsi->dev, "couldn't find panel port\n");
		return -ENXIO;
	}

	/* check other endpoints */
	for_each_endpoint_of_node(remote_port, ep) {
		struct device_node *np = of_graph_get_remote_port_parent(ep);

		if (!np)
			continue;

		idx = of_property_read_bool(np, "clock-master");

		/*
		 * Either master or slave already defined, drop refcnt
		 * but catch errors only after the full loop.
		 */
		if (ctrls[idx])
			of_node_put(np);
		else
			ctrls[idx] = np;

		num++;
	}
	of_node_put(remote_port);

	if (num > 2) {
		DRM_DEV_ERROR(dsi->dev, "too many dsi devices linked\n");
		ret = -EINVAL;
		goto cleanup;
	}

	/* nothing to do */
	if (num < 1) {
		ret = 0;
		goto cleanup;
	}

	if (!ctrls[1]) {
		DRM_DEV_ERROR(dsi->dev, "no master defined in dual-dsi\n");
		ret = -ENODEV;
		goto cleanup;
	}

	if (!ctrls[0]) {
		DRM_DEV_ERROR(dsi->dev, "no slave defined in dual-dsi\n");
		ret = -ENODEV;
		goto cleanup;
	}

	*master = ctrls[1];
	*slave = ctrls[0];

	return 1;

cleanup:

	for (idx = 0; idx < 2; idx++)
		if (ctrls[idx])
			of_node_put(ctrls[idx]);
	return ret;
}

Quoted 230 lines

Quoted 230 lines

So right now I'm operating with a devicetree like

&mipi_dsi {

          status = "okay";
          clock-master;
          
          ports {
          
                  mipi_out: port@1 {
                  
                          reg = <1>;
                          
                          mipi_out_panel: endpoint {
                          
                                  remote-endpoint = <&mipi_in_panel>;
                          
                          };
                  
                  };
          
          };
          
          mipi_panel: panel@0 {
		
		  compatible = "innolux,p097pfg";
		
                  reg = <0>;
                  backlight = <&backlight>;
                  enable-gpios = <&gpio4 25 GPIO_ACTIVE_HIGH>;
                  pinctrl-names = "default";
                  pinctrl-0 = <&display_rst_l>;
                  
                  port {
                  
                          #address-cells = <1>;
                          #size-cells = <0>;
                          
                          mipi_in_panel: endpoint@0 {
                          
                                  reg = <0>;
                                  remote-endpoint = <&mipi_out_panel>;
                          
                          };
                          
                          mipi1_in_panel: endpoint@1 {
                          
                                  reg = <1>;
                                  remote-endpoint = <&mipi1_out_panel>;
                          
                          };
                  
                  };
          
          };

};

&mipi_dsi1 {

          status = "okay";
          
          ports {
          
                  mipi1_out: port@1 {
                  
                          reg = <1>;
                          
                          mipi1_out_panel: endpoint {
                          
                                  remote-endpoint = <&mipi1_in_panel>;
                          
                          };
                  
                  };
          
          };

};

I guess it is a matter of preference on what reflects the hardware best, so maybe that's Robs call?

Heiko

Quoted 143 lines

Quoted 143 lines

Quoted 143 lines

---
v2:
- Specify that clock-master is a boolean property.
- Drop/add unit-address and #*-cells where applicable.

    .../devicetree/bindings/display/mipi-dsi-bus.txt   | 80
    ++++++++++++++++++++++ 1 file changed, 80 insertions(+)

(quoted diff) +0 -0

b/Documentation/devicetree/bindings/display/mipi-dsi-bus.txt index 94fb72cb916f..7a3abbedb3fa 100644

Documentation/devicetree/bindings/display/mipi-dsi-bus.txt +2 -0

@@ -29,6 +29,13 @@ Required properties:
    - #size-cells: Should be 0. There are cases where it makes sense to
    use
    a
 
      different value here. See below.
 
+Optional properties:
+- clock-master: boolean. Should be enabled if the host is being used

in
+ conjunction with another DSI host to drive the same peripheral. Hardware
+ supporting such a configuration generally requires the data on both the busses + to be driven by the same clock. Only the DSI host instance
controlling this + clock should contain this property.

Documentation/devicetree/bindings/display/mipi-dsi-bus.txt +1 -0

+
 
    DSI peripheral
    ==============

@@ -62,6 +69,16 @@ primary control bus, but are also connected to a DSI

bus (mostly for the data>>

path). Connections between such peripherals and a DSI host can be represented using the graph bindings [1], [2].

Documentation/devicetree/bindings/display/mipi-dsi-bus.txt +4 -0

+Peripherals that support dual channel DSI
+-----------------------------------------
+
+Peripherals with higher bandwidth requirements can be connected to 2

DSI
+busses. Each DSI bus/channel drives some portion of the pixel data (generally +left/right half of each line of the display, or even/odd lines of the display). +The graph bindings should be used to represent the multiple DSI busses that are +connected to this peripheral. Each DSI
host's output endpoint can be linked to +an input endpoint of the DSI peripheral.

Documentation/devicetree/bindings/display/mipi-dsi-bus.txt +2 -0

+
 
    [1] Documentation/devicetree/bindings/graph.txt
    [2] Documentation/devicetree/bindings/media/video-interfaces.txt

@@ -71,6 +88,8 @@ Examples
 
      with different virtual channel configurations.
 
    - (4) is an example of a peripheral on a I2C control bus connected
    with
    to
 
      a DSI host using of-graph bindings.
 
+- (5) is an example of 2 DSI hosts driving a dual-channel DSI

peripheral,

Documentation/devicetree/bindings/display/mipi-dsi-bus.txt +60 -0

+  which uses I2C as its primary control bus.
 
    1)
 
    	dsi-host {

@@ -153,3 +172,64 @@ Examples
 
    			};
 
    		};
 
    	};
 
+
+5)
+	i2c-host {
+		dsi-bridge@35 {
+			compatible = "...";
+			reg = <0x35>;
+
+			ports {
+				#address-cells = <1>;
+				#size-cells = <0>;
+
+				port@0 {
+					reg = <0>;
+					dsi0_in: endpoint {
+						remote-endpoint = <&dsi0_out>;
+					};
+				};
+
+				port@1 {
+					reg = <1>;
+					dsi1_in: endpoint {
+						remote-endpoint = <&dsi1_out>;
+					};
+				};
+			};
+		};
+	};
+
+	dsi0-host {
+		...
+
+		/*
+		 * this DSI instance drives the clock for both the host
+		 * controllers
+		 */
+		clock-master;
+
+		ports {
+			...
+
+			port {
+				dsi0_out: endpoint {
+					remote-endpoint = <&dsi0_in>;
+				};
+			};
+		};
+	};
+
+	dsi1-host {
+		...
+
+		ports {
+			...
+
+			port {
+				dsi1_out: endpoint {
+					remote-endpoint = <&dsi1_in>;
+				};
+			};

Quoted 8 lines

Quoted 8 lines

Quoted 8 lines

Quoted 8 lines

Documentation/devicetree/bindings/display/mipi-dsi-bus.txt

Quoted 2 lines

Documentation/devicetree/bindings/display/mipi-dsi-bus.txt +2 -0

+		};
+	};

--
To unsubscribe from this list: send the line "unsubscribe linux-arm-msm" in
the body of a message to [email protected]
More majordomo info at http://vger.kernel.org/majordomo-info.html

--
To unsubscribe from this list: send the line "unsubscribe linux-arm-msm" in the body of a message to [email protected]
More majordomo info at http://vger.kernel.org/majordomo-info.html

Quoted 44 lines

Hi Archit,

Am Mittwoch, 6. Juni 2018, 12:21:16 CEST schrieb Archit Taneja:

Quoted 41 lines · Heiko Stübner

Quoted 23 lines

Am Mittwoch, 6. Juni 2018, 07:59:29 CEST schrieb Archit Taneja:

Quoted 16 lines · Heiko Stuebner

Quoted 15 lines

Am Donnerstag, 18. Januar 2018, 05:53:55 CEST schrieb Archit Taneja:

Quoted 6 lines

Add binding info for peripherals that support dual-channel DSI. Add corresponding optional bindings for DSI host controllers that may be configured in this mode. Add an example of an I2C controlled device operating in dual-channel DSI mode.

Signed-off-by Archit Taneja <[email protected]>

Looks like a great solution for that problem, so

Reviewed-by Heiko Stuebner <[email protected]>

As I'm looking into that for my rk3399-scarlet device right now and couldn't find this patchset in the kernel yet, is it planned to merge or refresh these binding changes or were problems encountered.

Quoted 5 lines

Quoted 1 line

At least an Ack/Review from Rob seems to be missing.

I forgot about these patches. Rob had reviewed the first one in the set the second one still needed an Ack. I'll post a v3 that adds the Reviewed-bys and fixes a small typo.

Quoted 6 lines

very nice ... because it looks like yesterday I managed to make the Rockchip dsi work in dual mode following this.

But one question came up, do you really want two input ports on the panel side? I.e. hardware-wise, I guess the panel will have one 8-lane or so input thatonly gets split up on the soc side onto 2 dsi controllers?

I think all dual DSI panels actually have 2 DSI controllers/parsers within them, one on each port. The MIPI DSI spec doesn't support 8 lanes. Also, the pixels coming out of the host are distributed among the lanes differently than what would have been the case with a 'theoretical' 8 lane receiver.

Other than that, some dual DSI panels only accept DSI commands on the 'master' port, where as others expect the same command to be sent across both the ports.

Quoted 8 lines

Quoted 4 lines

Therefore, I think it's better to represent dual DSI panels having 2 DSI input ports.

Your DT looks good to me.

Hmm, that doesn't match up then ;-) ... as my dt uses 2 endpoints in one port for the dsi-links.

Sorry, I didn't notice you'd created two endpoints within a single port.

I don't think I'm particular about 2 ports vs 1 port with 2 endpoints. They both seem okay to me as long as we follow it consistently. I'm myself not 100% sure of how to figure where one should prefer endpoints over ports. Maybe someone more familiar with the of graph bindings could comment here.

Quoted 8 lines

The issue I see with using ports and not endpoints for dual-dsi links is with distinguishing between input and output ports.

For a panel that's easy, as you every port will be an input port and if you have 2, it's supposed dual-dsi. But for example I guess there might exist some dual-dsi-to-something bridges, where you would end up with say 3 (or even more) ports ... two dual-dsi inputs and 1 or more outputs.

Okay, I get your point here. Although, even if the remote device had exactly 2 ports. Is it safe to assume that port #0 is an input port and port #1 is an output port? Is that the norm?

I've at least seen one device (toshiba,tc358767 bridge) that can actually take either DPI as input or DSI based on how you configure it. There are 2 input ports here, port #0 for DSI and port #1 for DPI. Would it have made sense here to have a single port and 2 endpoints here too?

Quoted 3 lines

While the following argument might not be 100% valid from a dt-purity standpoint implementing this might get hairy for _any_ operating system, as you will need each panel/bridge to tell what the ports are used for.

Yeah.

Quoted 4 lines

I.e. in my endpoint based mapping, right now I have this nice and generic WIP function to parse the of_graph and get the master+slave nodes:

Quoted 2 lines

https://github.com/mmind/linux-rockchip/blob/tmp/rk3399-gru-bob-scarlet/drivers/gpu/drm/rockchip/dw-mipi-dsi-rockchip.c#L697
[0]

I'd tried out something locally before posting this patch, I don't have the code for it, but I can describe the steps I took. This code expects the panel/bridge to have 2 input ports.

1. DSI0 host driver looks up its output port, gets the remote endpoint, and get this endpoint's parent (using of_graph_get_remote_port_parent()). Keeps the parent device node in a temp variable.

2. DSI1 host driver does the same thing.

3. DSI0 and DSI1 check whether their outputs are connected to the same device. If so, they're in dual DSI mode. If not, they are operating independently.

The positive of this approach is that we don't need to make any assumptions about the panel/bridge's port numbers, which is great. The negative is that our DSI controller instances now need to query each other, which can be messy, but not too hard to implement.

I think the choice finally boils down to what makes more sense w.r.t representing the HW correctly. We'd need Rob's comment on that.

Thanks,
Archit

Quoted 339 lines

So I guess my proposal would be to have one port for inputs and one port for outputs for dsi peripherals, with possibly multiple endpoints in each.

Heiko

[0] github seems to have reliability problems, so for reference my parsing function:

static int dw_mipi_dsi_is_dual(struct dw_mipi_dsi_rockchip *dsi,
struct device_node **master, struct device_node **slave)

{
	struct device_node *local_ep, *remote_port, *ep;
	struct device_node *ctrls[2] = { NULL, NULL };
	int num = 0, ret = 0, idx;

	/* get local panel endpoint of the dsi controller */
	local_ep = of_graph_get_endpoint_by_regs(dsi->dev->of_node, 1, 0);
	if (!local_ep) {
		DRM_DEV_ERROR(dsi->dev, "couldn't find local panel endpoint\n");
		return -ENXIO;
	}

	/* get panel port */
	remote_port = of_graph_get_remote_port_parent(local_ep);
	of_node_put(local_ep);
	if (!remote_port) {
		DRM_DEV_ERROR(dsi->dev, "couldn't find panel port\n");
		return -ENXIO;
	}

	/* check other endpoints */
	for_each_endpoint_of_node(remote_port, ep) {
		struct device_node *np = of_graph_get_remote_port_parent(ep);

		if (!np)
			continue;

		idx = of_property_read_bool(np, "clock-master");

		/*
		 * Either master or slave already defined, drop refcnt
		 * but catch errors only after the full loop.
		 */
		if (ctrls[idx])
			of_node_put(np);
		else
			ctrls[idx] = np;

		num++;
	}
	of_node_put(remote_port);

	if (num > 2) {
		DRM_DEV_ERROR(dsi->dev, "too many dsi devices linked\n");
		ret = -EINVAL;
		goto cleanup;
	}

	/* nothing to do */
	if (num < 1) {
		ret = 0;
		goto cleanup;
	}

	if (!ctrls[1]) {
		DRM_DEV_ERROR(dsi->dev, "no master defined in dual-dsi\n");
		ret = -ENODEV;
		goto cleanup;
	}

	if (!ctrls[0]) {
		DRM_DEV_ERROR(dsi->dev, "no slave defined in dual-dsi\n");
		ret = -ENODEV;
		goto cleanup;
	}

	*master = ctrls[1];
	*slave = ctrls[0];

	return 1;

cleanup:

	for (idx = 0; idx < 2; idx++)
		if (ctrls[idx])
			of_node_put(ctrls[idx]);
	return ret;
}

Quoted 230 lines

Quoted 230 lines

So right now I'm operating with a devicetree like

&mipi_dsi {

          status = "okay";
          clock-master;
          
          ports {
          
                  mipi_out: port@1 {
                  
                          reg = <1>;
                          
                          mipi_out_panel: endpoint {
                          
                                  remote-endpoint = <&mipi_in_panel>;
                          
                          };
                  
                  };
          
          };
          
          mipi_panel: panel@0 {
		
		  compatible = "innolux,p097pfg";
		
                  reg = <0>;
                  backlight = <&backlight>;
                  enable-gpios = <&gpio4 25 GPIO_ACTIVE_HIGH>;
                  pinctrl-names = "default";
                  pinctrl-0 = <&display_rst_l>;
                  
                  port {
                  
                          #address-cells = <1>;
                          #size-cells = <0>;
                          
                          mipi_in_panel: endpoint@0 {
                          
                                  reg = <0>;
                                  remote-endpoint = <&mipi_out_panel>;
                          
                          };
                          
                          mipi1_in_panel: endpoint@1 {
                          
                                  reg = <1>;
                                  remote-endpoint = <&mipi1_out_panel>;
                          
                          };
                  
                  };
          
          };

};

&mipi_dsi1 {

          status = "okay";
          
          ports {
          
                  mipi1_out: port@1 {
                  
                          reg = <1>;
                          
                          mipi1_out_panel: endpoint {
                          
                                  remote-endpoint = <&mipi1_in_panel>;
                          
                          };
                  
                  };
          
          };

};

I guess it is a matter of preference on what reflects the hardware best, so maybe that's Robs call?

Heiko

Quoted 143 lines

Quoted 143 lines

Quoted 143 lines

---
v2:
- Specify that clock-master is a boolean property.
- Drop/add unit-address and #*-cells where applicable.

    .../devicetree/bindings/display/mipi-dsi-bus.txt   | 80
    ++++++++++++++++++++++ 1 file changed, 80 insertions(+)

(quoted diff) +0 -0

b/Documentation/devicetree/bindings/display/mipi-dsi-bus.txt index 94fb72cb916f..7a3abbedb3fa 100644

Documentation/devicetree/bindings/display/mipi-dsi-bus.txt +2 -0

@@ -29,6 +29,13 @@ Required properties:
    - #size-cells: Should be 0. There are cases where it makes sense to
    use
    a
 
      different value here. See below.
 
+Optional properties:
+- clock-master: boolean. Should be enabled if the host is being used

in
+ conjunction with another DSI host to drive the same peripheral. Hardware
+ supporting such a configuration generally requires the data on both the busses + to be driven by the same clock. Only the DSI host instance
controlling this + clock should contain this property.

Documentation/devicetree/bindings/display/mipi-dsi-bus.txt +1 -0

+
 
    DSI peripheral
    ==============

@@ -62,6 +69,16 @@ primary control bus, but are also connected to a DSI

bus (mostly for the data>>

path). Connections between such peripherals and a DSI host can be represented using the graph bindings [1], [2].

Documentation/devicetree/bindings/display/mipi-dsi-bus.txt +4 -0

+Peripherals that support dual channel DSI
+-----------------------------------------
+
+Peripherals with higher bandwidth requirements can be connected to 2

DSI
+busses. Each DSI bus/channel drives some portion of the pixel data (generally +left/right half of each line of the display, or even/odd lines of the display). +The graph bindings should be used to represent the multiple DSI busses that are +connected to this peripheral. Each DSI
host's output endpoint can be linked to +an input endpoint of the DSI peripheral.

Documentation/devicetree/bindings/display/mipi-dsi-bus.txt +2 -0

+
 
    [1] Documentation/devicetree/bindings/graph.txt
    [2] Documentation/devicetree/bindings/media/video-interfaces.txt

@@ -71,6 +88,8 @@ Examples
 
      with different virtual channel configurations.
 
    - (4) is an example of a peripheral on a I2C control bus connected
    with
    to
 
      a DSI host using of-graph bindings.
 
+- (5) is an example of 2 DSI hosts driving a dual-channel DSI

peripheral,

Documentation/devicetree/bindings/display/mipi-dsi-bus.txt +60 -0

+  which uses I2C as its primary control bus.
 
    1)
 
    	dsi-host {

@@ -153,3 +172,64 @@ Examples
 
    			};
 
    		};
 
    	};
 
+
+5)
+	i2c-host {
+		dsi-bridge@35 {
+			compatible = "...";
+			reg = <0x35>;
+
+			ports {
+				#address-cells = <1>;
+				#size-cells = <0>;
+
+				port@0 {
+					reg = <0>;
+					dsi0_in: endpoint {
+						remote-endpoint = <&dsi0_out>;
+					};
+				};
+
+				port@1 {
+					reg = <1>;
+					dsi1_in: endpoint {
+						remote-endpoint = <&dsi1_out>;
+					};
+				};
+			};
+		};
+	};
+
+	dsi0-host {
+		...
+
+		/*
+		 * this DSI instance drives the clock for both the host
+		 * controllers
+		 */
+		clock-master;
+
+		ports {
+			...
+
+			port {
+				dsi0_out: endpoint {
+					remote-endpoint = <&dsi0_in>;
+				};
+			};
+		};
+	};
+
+	dsi1-host {
+		...
+
+		ports {
+			...
+
+			port {
+				dsi1_out: endpoint {
+					remote-endpoint = <&dsi1_in>;
+				};
+			};

Quoted 8 lines

Quoted 8 lines

Quoted 8 lines

Quoted 8 lines

Documentation/devicetree/bindings/display/mipi-dsi-bus.txt

Quoted 2 lines

Documentation/devicetree/bindings/display/mipi-dsi-bus.txt +2 -0

+		};
+	};

--
To unsubscribe from this list: send the line "unsubscribe linux-arm-msm" in
the body of a message to [email protected]
More majordomo info at http://vger.kernel.org/majordomo-info.html

--
To unsubscribe from this list: send the line "unsubscribe linux-arm-msm" in the body of a message to [email protected]
More majordomo info at http://vger.kernel.org/majordomo-info.html

Am Mittwoch, 6. Juni 2018, 18:07:46 CEST schrieb Archit Taneja:

Quoted 55 lines · Heiko Stübner

Quoted 40 lines

Hi Archit,

Am Mittwoch, 6. Juni 2018, 12:21:16 CEST schrieb Archit Taneja:

Quoted 37 lines · Heiko Stübner

Quoted 20 lines

Am Mittwoch, 6. Juni 2018, 07:59:29 CEST schrieb Archit Taneja:

Quoted 11 lines · Heiko Stuebner

Quoted 10 lines

Am Donnerstag, 18. Januar 2018, 05:53:55 CEST schrieb Archit Taneja:

Quoted 6 lines

Add binding info for peripherals that support dual-channel DSI. Add corresponding optional bindings for DSI host controllers that may be configured in this mode. Add an example of an I2C controlled device operating in dual-channel DSI mode.

Signed-off-by Archit Taneja <[email protected]>

Looks like a great solution for that problem, so

Reviewed-by Heiko Stuebner <[email protected]>

Quoted 8 lines

Quoted 5 lines

As I'm looking into that for my rk3399-scarlet device right now and couldn't find this patchset in the kernel yet, is it planned to merge or refresh these binding changes or were problems encountered.

At least an Ack/Review from Rob seems to be missing.

I forgot about these patches. Rob had reviewed the first one in the set the second one still needed an Ack. I'll post a v3 that adds the Reviewed-bys and fixes a small typo.

Quoted 6 lines

very nice ... because it looks like yesterday I managed to make the Rockchip dsi work in dual mode following this.

But one question came up, do you really want two input ports on the panel side? I.e. hardware-wise, I guess the panel will have one 8-lane or so input thatonly gets split up on the soc side onto 2 dsi controllers?

I think all dual DSI panels actually have 2 DSI controllers/parsers within them, one on each port. The MIPI DSI spec doesn't support 8 lanes. Also, the pixels coming out of the host are distributed among the lanes differently than what would have been the case with a 'theoretical' 8 lane receiver.

Other than that, some dual DSI panels only accept DSI commands on the 'master' port, where as others expect the same command to be sent across both the ports.

Quoted 7 lines

Quoted 4 lines

Therefore, I think it's better to represent dual DSI panels having 2 DSI input ports.

Your DT looks good to me.

Hmm, that doesn't match up then ;-) ... as my dt uses 2 endpoints in one port for the dsi-links.

Sorry, I didn't notice you'd created two endpoints within a single port.

I don't think I'm particular about 2 ports vs 1 port with 2 endpoints. They both seem okay to me as long as we follow it consistently. I'm myself not 100% sure of how to figure where one should prefer endpoints over ports. Maybe someone more familiar with the of graph bindings could comment here.

Quoted 10 lines

Quoted 10 lines

The issue I see with using ports and not endpoints for dual-dsi links is with distinguishing between input and output ports.

For a panel that's easy, as you every port will be an input port and if you have 2, it's supposed dual-dsi. But for example I guess there might exist some dual-dsi-to-something bridges, where you would end up with say 3 (or even more) ports ... two dual-dsi inputs and 1 or more outputs.

Quoted 3 lines

Okay, I get your point here. Although, even if the remote device had exactly 2 ports. Is it safe to assume that port #0 is an input port and port #1 is an output port? Is that the norm?

I don't think that anything like that is specified anywhere, so you cannot assume anything about what a port contains.

Quoted 6 lines

I've at least seen one device (toshiba,tc358767 bridge) that can actually take either DPI as input or DSI based on how you configure it. There are 2 input ports here, port #0 for DSI and port #1 for DPI. Would it have made sense here to have a single port and 2 endpoints here too?

Nope, I guess having a port for DPI input, one port for DSI input makes quite a lot of sense. And then you can have the input-specifics living in the endpoints like dual links or so.

Quoted 4 lines

Quoted 3 lines

While the following argument might not be 100% valid from a dt-purity standpoint implementing this might get hairy for _any_ operating system, as you will need each panel/bridge to tell what the ports are used for.

Yeah.

Quoted 4 lines

Quoted 4 lines

I.e. in my endpoint based mapping, right now I have this nice and generic WIP function to parse the of_graph and get the master+slave nodes:

Quoted 5 lines

Quoted 2 lines

https://github.com/mmind/linux-rockchip/blob/tmp/rk3399-gru-bob-scarlet/drivers/gpu/drm/rockchip/dw-mipi-dsi-rockchip.c#L697
[0]

I'd tried out something locally before posting this patch, I don't have the code for it, but I can describe the steps I took. This code expects the panel/bridge to have 2 input ports.

Which again would be very much panel-specific as you cannot assume to much about the ports.

Quoted 15 lines

1. DSI0 host driver looks up its output port, gets the remote endpoint, and get this endpoint's parent (using of_graph_get_remote_port_parent()). Keeps the parent device node in a temp variable.

2. DSI1 host driver does the same thing.

3. DSI0 and DSI1 check whether their outputs are connected to the same device. If so, they're in dual DSI mode. If not, they are operating independently.

The positive of this approach is that we don't need to make any assumptions about the panel/bridge's port numbers, which is great. The negative is that our DSI controller instances now need to query each other, which can be messy, but not too hard to implement.

Quoted 3 lines

I think the choice finally boils down to what makes more sense w.r.t representing the HW correctly. We'd need Rob's comment on that.

Taking your DPI+DSI example from above actually shows quite nicely how ports+endpoints could play together.

panel-or-bridge@0 {
	compatible = "something,something";
	reg = <0>;

	ports {
		/* DPI input */
		port@0 {
			endpoint {
				remote-endpoint = <&dpi_out>;
			};
		};

		/* DSI input */
		port@1 {
			/* from first dsi controller */
			endpoint@0 {
				remote-endpoint = <&dsi0_out>;
			};

			/* from second dsi controller */
			endpoint@1 {
				remote-endpoint = <&dsi1_out>;
			};
		};

		/* another input, or an output for a bridge */
		port@2 {
			endpoint {
				remote-endpoint = <&somewhere>;
			}
		};
	};
};

But yeah, hopefully we can entice Rob for comments :-) .

Heiko

Quoted 340 lines

Thanks,
Archit

Quoted 336 lines

So I guess my proposal would be to have one port for inputs and one port for outputs for dsi peripherals, with possibly multiple endpoints in each.

Heiko

[0] github seems to have reliability problems, so for reference my parsing function:

static int dw_mipi_dsi_is_dual(struct dw_mipi_dsi_rockchip *dsi,
struct device_node **master, struct device_node **slave)

{
	struct device_node *local_ep, *remote_port, *ep;
	struct device_node *ctrls[2] = { NULL, NULL };
	int num = 0, ret = 0, idx;

	/* get local panel endpoint of the dsi controller */
	local_ep = of_graph_get_endpoint_by_regs(dsi->dev->of_node, 1, 0);
	if (!local_ep) {
		DRM_DEV_ERROR(dsi->dev, "couldn't find local panel endpoint\n");
		return -ENXIO;
	}

	/* get panel port */
	remote_port = of_graph_get_remote_port_parent(local_ep);
	of_node_put(local_ep);
	if (!remote_port) {
		DRM_DEV_ERROR(dsi->dev, "couldn't find panel port\n");
		return -ENXIO;
	}

	/* check other endpoints */
	for_each_endpoint_of_node(remote_port, ep) {
		struct device_node *np = of_graph_get_remote_port_parent(ep);

		if (!np)
			continue;

		idx = of_property_read_bool(np, "clock-master");

		/*
		 * Either master or slave already defined, drop refcnt
		 * but catch errors only after the full loop.
		 */
		if (ctrls[idx])
			of_node_put(np);
		else
			ctrls[idx] = np;

		num++;
	}
	of_node_put(remote_port);

	if (num > 2) {
		DRM_DEV_ERROR(dsi->dev, "too many dsi devices linked\n");
		ret = -EINVAL;
		goto cleanup;
	}

	/* nothing to do */
	if (num < 1) {
		ret = 0;
		goto cleanup;
	}

	if (!ctrls[1]) {
		DRM_DEV_ERROR(dsi->dev, "no master defined in dual-dsi\n");
		ret = -ENODEV;
		goto cleanup;
	}

	if (!ctrls[0]) {
		DRM_DEV_ERROR(dsi->dev, "no slave defined in dual-dsi\n");
		ret = -ENODEV;
		goto cleanup;
	}

	*master = ctrls[1];
	*slave = ctrls[0];

	return 1;

cleanup:

	for (idx = 0; idx < 2; idx++)
		if (ctrls[idx])
			of_node_put(ctrls[idx]);
	return ret;
}

Quoted 229 lines

Quoted 229 lines

So right now I'm operating with a devicetree like

&mipi_dsi {

          status = "okay";
          clock-master;
          
          ports {
          
                  mipi_out: port@1 {
                  
                          reg = <1>;
                          
                          mipi_out_panel: endpoint {
                          
                                  remote-endpoint = <&mipi_in_panel>;
                          
                          };
                  
                  };
          
          };
          
          mipi_panel: panel@0 {
		
		  compatible = "innolux,p097pfg";
		
                  reg = <0>;
                  backlight = <&backlight>;
                  enable-gpios = <&gpio4 25 GPIO_ACTIVE_HIGH>;
                  pinctrl-names = "default";
                  pinctrl-0 = <&display_rst_l>;
                  
                  port {
                  
                          #address-cells = <1>;
                          #size-cells = <0>;
                          
                          mipi_in_panel: endpoint@0 {
                          
                                  reg = <0>;
                                  remote-endpoint = <&mipi_out_panel>;
                          
                          };
                          
                          mipi1_in_panel: endpoint@1 {
                          
                                  reg = <1>;
                                  remote-endpoint = <&mipi1_out_panel>;
                          
                          };
                  
                  };
          
          };

};

&mipi_dsi1 {

          status = "okay";
          
          ports {
          
                  mipi1_out: port@1 {
                  
                          reg = <1>;
                          
                          mipi1_out_panel: endpoint {
                          
                                  remote-endpoint = <&mipi1_in_panel>;
                          
                          };
                  
                  };
          
          };

};

I guess it is a matter of preference on what reflects the hardware best, so maybe that's Robs call?

Heiko

Quoted 142 lines

Quoted 142 lines

Quoted 142 lines

---
v2:
- Specify that clock-master is a boolean property.
- Drop/add unit-address and #*-cells where applicable.

    .../devicetree/bindings/display/mipi-dsi-bus.txt   | 80
    ++++++++++++++++++++++ 1 file changed, 80 insertions(+)

(quoted diff) +0 -0

b/Documentation/devicetree/bindings/display/mipi-dsi-bus.txt index 94fb72cb916f..7a3abbedb3fa 100644

Documentation/devicetree/bindings/display/mipi-dsi-bus.txt +2 -0

@@ -29,6 +29,13 @@ Required properties:
    - #size-cells: Should be 0. There are cases where it makes sense to
    use
    a
 
      different value here. See below.
 
+Optional properties:
+- clock-master: boolean. Should be enabled if the host is being used

in
+ conjunction with another DSI host to drive the same peripheral. Hardware
+ supporting such a configuration generally requires the data on both the busses + to be driven by the same clock. Only the DSI host instance
controlling this + clock should contain this property.

Documentation/devicetree/bindings/display/mipi-dsi-bus.txt +1 -0

+
 
    DSI peripheral
    ==============

@@ -62,6 +69,16 @@ primary control bus, but are also connected to a DSI

bus (mostly for the data>>

path). Connections between such peripherals and a DSI host can be represented using the graph bindings [1], [2].

Documentation/devicetree/bindings/display/mipi-dsi-bus.txt +4 -0

+Peripherals that support dual channel DSI
+-----------------------------------------
+
+Peripherals with higher bandwidth requirements can be connected to 2

DSI
+busses. Each DSI bus/channel drives some portion of the pixel data (generally +left/right half of each line of the display, or even/odd lines of the display). +The graph bindings should be used to represent the multiple DSI busses that are +connected to this peripheral. Each DSI
host's output endpoint can be linked to +an input endpoint of the DSI peripheral.

Documentation/devicetree/bindings/display/mipi-dsi-bus.txt +2 -0

+
 
    [1] Documentation/devicetree/bindings/graph.txt
    [2] Documentation/devicetree/bindings/media/video-interfaces.txt

@@ -71,6 +88,8 @@ Examples
 
      with different virtual channel configurations.
 
    - (4) is an example of a peripheral on a I2C control bus connected
    with
    to
 
      a DSI host using of-graph bindings.
 
+- (5) is an example of 2 DSI hosts driving a dual-channel DSI

peripheral,

Documentation/devicetree/bindings/display/mipi-dsi-bus.txt +59 -0

+  which uses I2C as its primary control bus.
 
    1)
 
    	dsi-host {

@@ -153,3 +172,64 @@ Examples
 
    			};
 
    		};
 
    	};
 
+
+5)
+	i2c-host {
+		dsi-bridge@35 {
+			compatible = "...";
+			reg = <0x35>;
+
+			ports {
+				#address-cells = <1>;
+				#size-cells = <0>;
+
+				port@0 {
+					reg = <0>;
+					dsi0_in: endpoint {
+						remote-endpoint = <&dsi0_out>;
+					};
+				};
+
+				port@1 {
+					reg = <1>;
+					dsi1_in: endpoint {
+						remote-endpoint = <&dsi1_out>;
+					};
+				};
+			};
+		};
+	};
+
+	dsi0-host {
+		...
+
+		/*
+		 * this DSI instance drives the clock for both the host
+		 * controllers
+		 */
+		clock-master;
+
+		ports {
+			...
+
+			port {
+				dsi0_out: endpoint {
+					remote-endpoint = <&dsi0_in>;
+				};
+			};
+		};
+	};
+
+	dsi1-host {
+		...
+
+		ports {
+			...
+
+			port {
+				dsi1_out: endpoint {
+					remote-endpoint = <&dsi1_in>;
+				};

Quoted 8 lines

Quoted 8 lines

Quoted 8 lines

Quoted 8 lines

Documentation/devicetree/bindings/display/mipi-dsi-bus.txt

Quoted 3 lines

Documentation/devicetree/bindings/display/mipi-dsi-bus.txt +3 -0

+			};
+		};
+	};

--
To unsubscribe from this list: send the line "unsubscribe linux-arm-msm" in
the body of a message to [email protected]
More majordomo info at http://vger.kernel.org/majordomo-info.html

--
To unsubscribe from this list: send the line "unsubscribe linux-arm-msm" in the body of a message to [email protected]
More majordomo info at http://vger.kernel.org/majordomo-info.html

_______________________________________________
dri-devel mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/dri-devel

Quoted 63 lines

Am Mittwoch, 6. Juni 2018, 18:07:46 CEST schrieb Archit Taneja:

Quoted 62 lines · Heiko Stübner

Quoted 46 lines

Hi Archit,

Am Mittwoch, 6. Juni 2018, 12:21:16 CEST schrieb Archit Taneja:

Quoted 43 lines · Heiko Stübner

Quoted 23 lines

Am Mittwoch, 6. Juni 2018, 07:59:29 CEST schrieb Archit Taneja:

Quoted 16 lines · Heiko Stuebner

Quoted 15 lines

Am Donnerstag, 18. Januar 2018, 05:53:55 CEST schrieb Archit Taneja:

Quoted 6 lines

Add binding info for peripherals that support dual-channel DSI. Add corresponding optional bindings for DSI host controllers that may be configured in this mode. Add an example of an I2C controlled device operating in dual-channel DSI mode.

Signed-off-by Archit Taneja <[email protected]>

Looks like a great solution for that problem, so

Reviewed-by Heiko Stuebner <[email protected]>

As I'm looking into that for my rk3399-scarlet device right now and couldn't find this patchset in the kernel yet, is it planned to merge or refresh these binding changes or were problems encountered.

Quoted 5 lines

Quoted 1 line

At least an Ack/Review from Rob seems to be missing.

I forgot about these patches. Rob had reviewed the first one in the set the second one still needed an Ack. I'll post a v3 that adds the Reviewed-bys and fixes a small typo.

Quoted 7 lines

very nice ... because it looks like yesterday I managed to make the Rockchip dsi work in dual mode following this.

But one question came up, do you really want two input ports on the panel
side? I.e. hardware-wise, I guess the panel will have one 8-lane or so input thatonly gets split up on the soc side onto 2 dsi controllers?

I think all dual DSI panels actually have 2 DSI controllers/parsers within them, one on each port. The MIPI DSI spec doesn't support 8 lanes. Also, the pixels coming out of the host are distributed among the lanes differently than what would have been the case with a 'theoretical' 8 lane receiver.

Other than that, some dual DSI panels only accept DSI commands on the 'master' port, where as others expect the same command to be sent across
both the ports.

Quoted 7 lines

Quoted 4 lines

Therefore, I think it's better to represent dual DSI panels having 2 DSI input ports.

Your DT looks good to me.

Hmm, that doesn't match up then ;-) ... as my dt uses 2 endpoints in one port for the dsi-links.

Sorry, I didn't notice you'd created two endpoints within a single port.

I don't think I'm particular about 2 ports vs 1 port with 2 endpoints. They both seem okay to me as long as we follow it consistently. I'm myself not 100% sure of how to figure where one should prefer endpoints over ports. Maybe someone more familiar with the of graph bindings could comment here.

Port should describe physical port, endpoint are used to describe multiple links to the same port. If the panel has two physical DSI interfaces it should use two ports.

Quoted 9 lines

Quoted 9 lines

Quoted 8 lines

The issue I see with using ports and not endpoints for dual-dsi links is with distinguishing between input and output ports.

For a panel that's easy, as you every port will be an input port and if you have 2, it's supposed dual-dsi. But for example I guess there might exist some dual-dsi-to-something bridges, where you would end up with say 3 (or even more) ports ... two dual-dsi inputs and 1 or more outputs.

Quoted 6 lines

Quoted 3 lines

Okay, I get your point here. Although, even if the remote device had exactly 2 ports. Is it safe to assume that port #0 is an input port and port #1 is an output port? Is that the norm?

I don't think that anything like that is specified anywhere, so you cannot assume anything about what a port contains.

Ports are defined per device-node and port number assignment should be described in particular device bindings. So device driver knows well which functions should be assigned to which port. But it is private matter of the device/device driver. Different device bindings can assign different number for input and output ports, but since ports are private thing for devices it should not be a problem.

I do not see why do you have such issue, could you describe it more.

Quoted 9 lines

Quoted 4 lines

I've at least seen one device (toshiba,tc358767 bridge) that can actually take either DPI as input or DSI based on how you configure it. There are 2 input ports here, port #0 for DSI and port #1 for DPI. Would it have made sense here to have a single port and 2 endpoints here too?

Nope, I guess having a port for DPI input, one port for DSI input makes quite a lot of sense. And then you can have the input-specifics living in the endpoints like dual links or so.

Quoted 5 lines

Quoted 5 lines

Quoted 3 lines

While the following argument might not be 100% valid from a dt-purity standpoint implementing this might get hairy for _any_ operating system, as you will need each panel/bridge to tell what the ports are used for.

Yeah.

Each panel/bridge should know what his ports are used for, but to who and why you want to communicate it?

Quoted 7 lines

Quoted 7 lines

Quoted 6 lines

I.e. in my endpoint based mapping, right now I have this nice and generic
WIP function to parse the of_graph and get the master+slave nodes:

https://github.com/mmind/linux-rockchip/blob/tmp/rk3399-gru-bob-scarlet/
drivers/gpu/drm/rockchip/dw-mipi-dsi-rockchip.c#L697 [0]

Quoted 22 lines

Quoted 3 lines

I'd tried out something locally before posting this patch, I don't have the code for it, but I can describe the steps I took. This code expects the panel/bridge to have 2 input ports.

Which again would be very much panel-specific as you cannot assume to much about the ports.

Quoted 15 lines

1. DSI0 host driver looks up its output port, gets the remote endpoint, and get this endpoint's parent (using of_graph_get_remote_port_parent()). Keeps the parent device node in a temp variable.

2. DSI1 host driver does the same thing.

3. DSI0 and DSI1 check whether their outputs are connected to the same device. If so, they're in dual DSI mode. If not, they are operating independently.

The positive of this approach is that we don't need to make any assumptions about the panel/bridge's port numbers, which is great. The negative is that our DSI controller instances now need to query each other, which can be messy, but not too hard to implement.

I think detection of dual dsi and whole configuration thing of dual-dsi should be done using in-kernel frameworks (maybe new ones :) ). Could you describe necessary configuration steps to be performed, what info should be passed from who to who, etc. It would be easier to discuss.

Anyway I think there are rules which probably we should always (almost) obey:
1. one port per interface, if it is control port it could be probably omitted.
2. device drivers should only look into their private nodes, nodes pointed by remote-endpoints should be used only/mainly for identification of in-kernel objects exposed by in-kernel frameworks.

Am Donnerstag, 7. Juni 2018, 12:39:03 CEST schrieb Andrzej Hajda:

Quoted 63 lines · Heiko Stuebner

Quoted 61 lines

Am Mittwoch, 6. Juni 2018, 18:07:46 CEST schrieb Archit Taneja:

Quoted 60 lines · Heiko Stübner

Quoted 44 lines

Am Mittwoch, 6. Juni 2018, 12:21:16 CEST schrieb Archit Taneja:

Quoted 43 lines · Heiko Stübner

Quoted 23 lines

Am Mittwoch, 6. Juni 2018, 07:59:29 CEST schrieb Archit Taneja:

Quoted 16 lines · Heiko Stuebner

Quoted 15 lines

Am Donnerstag, 18. Januar 2018, 05:53:55 CEST schrieb Archit Taneja:

Quoted 6 lines

Add binding info for peripherals that support dual-channel DSI. Add corresponding optional bindings for DSI host controllers that may be configured in this mode. Add an example of an I2C controlled device operating in dual-channel DSI mode.

Signed-off-by Archit Taneja <[email protected]>

Looks like a great solution for that problem, so

Reviewed-by Heiko Stuebner <[email protected]>

As I'm looking into that for my rk3399-scarlet device right now and couldn't find this patchset in the kernel yet, is it planned to merge or refresh these binding changes or were problems encountered.

Quoted 5 lines

Quoted 1 line

At least an Ack/Review from Rob seems to be missing.

I forgot about these patches. Rob had reviewed the first one in the set the second one still needed an Ack. I'll post a v3 that adds the Reviewed-bys and fixes a small typo.

Quoted 7 lines

very nice ... because it looks like yesterday I managed to make the Rockchip dsi work in dual mode following this.

But one question came up, do you really want two input ports on the panel
side? I.e. hardware-wise, I guess the panel will have one 8-lane or so input thatonly gets split up on the soc side onto 2 dsi controllers?

I think all dual DSI panels actually have 2 DSI controllers/parsers within them, one on each port. The MIPI DSI spec doesn't support 8 lanes. Also, the pixels coming out of the host are distributed among the lanes differently than what would have been the case with a 'theoretical' 8 lane receiver.

Other than that, some dual DSI panels only accept DSI commands on the 'master' port, where as others expect the same command to be sent across
both the ports.

Quoted 7 lines

Quoted 4 lines

Therefore, I think it's better to represent dual DSI panels having 2 DSI input ports.

Your DT looks good to me.

Hmm, that doesn't match up then ;-) ... as my dt uses 2 endpoints in one port for the dsi-links.

Sorry, I didn't notice you'd created two endpoints within a single port.

I don't think I'm particular about 2 ports vs 1 port with 2 endpoints. They both seem okay to me as long as we follow it consistently. I'm myself not 100% sure of how to figure where one should prefer endpoints over ports. Maybe someone more familiar with the of graph bindings could comment here.

Quoted 3 lines

Port should describe physical port, endpoint are used to describe multiple links to the same port. If the panel has two physical DSI interfaces it should use two ports.

That again leaves a bit of space for interpretation ;-) .

A dual-dsi display is probably pretty useless with only one controller connected to it, as its 4 lanes cannot satisfy the required 8 lanes of the panel.

See [0] for current WIP panel addition.
[Was already on the lists but needs cleanup]

[0] https://github.com/mmind/linux-rockchip/commit/459bc1a1599377c2c5408724e82619a4602a953d#diff-5d6d6ddab4fd282102d23d2c02e835f8R381

Quoted 22 lines

Quoted 9 lines

Quoted 9 lines

Quoted 8 lines

The issue I see with using ports and not endpoints for dual-dsi links is with distinguishing between input and output ports.

For a panel that's easy, as you every port will be an input port and if you have 2, it's supposed dual-dsi. But for example I guess there might exist some dual-dsi-to-something bridges, where you would end up with say 3 (or even more) ports ... two dual-dsi inputs and 1 or more outputs.

Quoted 6 lines

Quoted 3 lines

Okay, I get your point here. Although, even if the remote device had exactly 2 ports. Is it safe to assume that port #0 is an input port and port #1 is an output port? Is that the norm?

I don't think that anything like that is specified anywhere, so you cannot assume anything about what a port contains.

Ports are defined per device-node and port number assignment should be described in particular device bindings. So device driver knows well which functions should be assigned to which port. But it is private matter of the device/device driver. Different device bindings can assign different number for input and output ports, but since ports are private thing for devices it should not be a problem.

Quoted 11 lines

I do not see why do you have such issue, could you describe it more.

Quoted 9 lines

Quoted 4 lines

I've at least seen one device (toshiba,tc358767 bridge) that can actually take either DPI as input or DSI based on how you configure it. There are 2 input ports here, port #0 for DSI and port #1 for DPI. Would it have made sense here to have a single port and 2 endpoints here too?

Nope, I guess having a port for DPI input, one port for DSI input makes quite a lot of sense. And then you can have the input-specifics living in the endpoints like dual links or so.

Quoted 8 lines

Quoted 5 lines

Quoted 5 lines

Quoted 3 lines

While the following argument might not be 100% valid from a dt-purity standpoint implementing this might get hairy for _any_ operating system, as you will need each panel/bridge to tell what the ports are used for.

Yeah.

Each panel/bridge should know what his ports are used for, but to who and why you want to communicate it?

I'll do that below :-)

Quoted 30 lines

Quoted 7 lines

Quoted 7 lines

Quoted 6 lines

I.e. in my endpoint based mapping, right now I have this nice and generic
WIP function to parse the of_graph and get the master+slave nodes:

https://github.com/mmind/linux-rockchip/blob/tmp/rk3399-gru-bob-scarlet/
drivers/gpu/drm/rockchip/dw-mipi-dsi-rockchip.c#L697 [0]

Quoted 22 lines

Quoted 3 lines

I'd tried out something locally before posting this patch, I don't have the code for it, but I can describe the steps I took. This code expects the panel/bridge to have 2 input ports.

Which again would be very much panel-specific as you cannot assume to much about the ports.

Quoted 15 lines

1. DSI0 host driver looks up its output port, gets the remote endpoint, and get this endpoint's parent (using of_graph_get_remote_port_parent()). Keeps the parent device node in a temp variable.

2. DSI1 host driver does the same thing.

3. DSI0 and DSI1 check whether their outputs are connected to the same device. If so, they're in dual DSI mode. If not, they are operating independently.

The positive of this approach is that we don't need to make any assumptions about the panel/bridge's port numbers, which is great. The negative is that our DSI controller instances now need to query each other, which can be messy, but not too hard to implement.

Quoted 4 lines

I think detection of dual dsi and whole configuration thing of dual-dsi should be done using in-kernel frameworks (maybe new ones :) ). Could you describe necessary configuration steps to be performed, what info should be passed from who to who, etc. It would be easier to discuss.

The base setup is the rk3399 soc, which has 2 completely separate dw-mipi-dsi controllers.

Gru-Scarlet (a ChromeOS tablet device) uses a high-res dual-DSI display (2500x1500 pixels or so) which is driven by both controllers together to achieve the required number of lanes to it.

So you end up with one (specific) controller being the master and the other being the slave, deferring to the master for most everything. As the controllers are somehow build to work together in that case via special soc-specific settings.

See [1] for current WIP code, I lifted out of the chromeos kernel and modified to not have that special rockchip,dual-channel =<&...>; property. That patch was on the lists already last year I think.

So each dw-mipi-dsi instance needs to at least know that it is in a master-slave-relationship and right now also needs access to the other driver instance. Doing the master->slave access should be fine in that special case, as the IP is the same type.

Also the crtc<->dsi interaction is quite a bit of handling-different between one crtc talking to 2 DSIs or 2 crtc talking to the 2 DSIs separately.

Hi,

I only have a little bit to add, but Heiko did solicit my opinion.

Quoted 64 lines

Am Donnerstag, 7. Juni 2018, 12:39:03 CEST schrieb Andrzej Hajda:

Quoted 63 lines · Heiko Stuebner

Quoted 61 lines

Am Mittwoch, 6. Juni 2018, 18:07:46 CEST schrieb Archit Taneja:

Quoted 60 lines · Heiko Stübner

Quoted 44 lines

Am Mittwoch, 6. Juni 2018, 12:21:16 CEST schrieb Archit Taneja:

Quoted 43 lines · Heiko Stübner

Quoted 23 lines

Am Mittwoch, 6. Juni 2018, 07:59:29 CEST schrieb Archit Taneja:

Quoted 16 lines · Heiko Stuebner

Quoted 15 lines

Am Donnerstag, 18. Januar 2018, 05:53:55 CEST schrieb Archit Taneja:

Quoted 6 lines

Add binding info for peripherals that support dual-channel DSI. Add corresponding optional bindings for DSI host controllers that may be configured in this mode. Add an example of an I2C controlled device operating in dual-channel DSI mode.

Signed-off-by Archit Taneja <[email protected]>

Looks like a great solution for that problem, so

Reviewed-by Heiko Stuebner <[email protected]>

As I'm looking into that for my rk3399-scarlet device right now and couldn't find this patchset in the kernel yet, is it planned to merge or refresh these binding changes or were problems encountered.

Quoted 5 lines

Quoted 1 line

At least an Ack/Review from Rob seems to be missing.

I forgot about these patches. Rob had reviewed the first one in the set the second one still needed an Ack. I'll post a v3 that adds the Reviewed-bys and fixes a small typo.

Quoted 7 lines

very nice ... because it looks like yesterday I managed to make the Rockchip dsi work in dual mode following this.

But one question came up, do you really want two input ports on the panel
side? I.e. hardware-wise, I guess the panel will have one 8-lane or so input thatonly gets split up on the soc side onto 2 dsi controllers?

I think all dual DSI panels actually have 2 DSI controllers/parsers within them, one on each port. The MIPI DSI spec doesn't support 8 lanes. Also, the pixels coming out of the host are distributed among the lanes differently than what would have been the case with a 'theoretical' 8 lane receiver.

Other than that, some dual DSI panels only accept DSI commands on the 'master' port, where as others expect the same command to be sent across
both the ports.

Quoted 7 lines

Quoted 4 lines

Therefore, I think it's better to represent dual DSI panels having 2 DSI input ports.

Your DT looks good to me.

Hmm, that doesn't match up then ;-) ... as my dt uses 2 endpoints in one port for the dsi-links.

Sorry, I didn't notice you'd created two endpoints within a single port.

I don't think I'm particular about 2 ports vs 1 port with 2 endpoints. They both seem okay to me as long as we follow it consistently. I'm myself not 100% sure of how to figure where one should prefer endpoints over ports. Maybe someone more familiar with the of graph bindings could comment here.

Quoted 5 lines

Quoted 3 lines

Port should describe physical port, endpoint are used to describe multiple links to the same port. If the panel has two physical DSI interfaces it should use two ports.

That again leaves a bit of space for interpretation ;-) .

Does it really? DSI maxes out at 4 lanes, so more than 4 lanes is clearly not a single standard port. Additionally, the code we're currently using in Chrome OS actually sends the same commands to each DSI separately (I'm not sure if it's actually required by the panels were using), so again, they are to some extent logically independent.

Quoted 3 lines

A dual-dsi display is probably pretty useless with only one controller connected to it, as its 4 lanes cannot satisfy the required 8 lanes of the panel.

Probably, although technically you can usually display on half the screen, if you have only have 1 working DSI.

Quoted 5 lines

See [0] for current WIP panel addition.
[Was already on the lists but needs cleanup]

Quoted 25 lines

[0] https://github.com/mmind/linux-rockchip/commit/459bc1a1599377c2c5408724e82619a4602a953d#diff-5d6d6ddab4fd282102d23d2c02e835f8R381

Quoted 22 lines

Quoted 9 lines

Quoted 9 lines

Quoted 8 lines

The issue I see with using ports and not endpoints for dual-dsi links is with distinguishing between input and output ports.

For a panel that's easy, as you every port will be an input port and if you have 2, it's supposed dual-dsi. But for example I guess there might exist some dual-dsi-to-something bridges, where you would end up with say 3 (or even more) ports ... two dual-dsi inputs and 1 or more outputs.

Quoted 6 lines

Quoted 3 lines

Okay, I get your point here. Although, even if the remote device had exactly 2 ports. Is it safe to assume that port #0 is an input port and port #1 is an output port? Is that the norm?

I don't think that anything like that is specified anywhere, so you cannot assume anything about what a port contains.

Ports are defined per device-node and port number assignment should be described in particular device bindings. So device driver knows well which functions should be assigned to which port. But it is private matter of the device/device driver. Different device bindings can assign different number for input and output ports, but since ports are private thing for devices it should not be a problem.

Agreed. And if Heiko needs a specific example close to home: all the Rockchip display bindings do this similarly; for example display/rockchip/analogix_dp-rockchip.txt describes exactly which port numbers and enpoints mean what:

- ports: there are 2 port nodes with endpoint definitions as defined in Documentation/devicetree/bindings/media/video-interfaces.txt. Port 0: contained 2 endpoints, connecting to the output of vop. Port 1: contained 1 endpoint, connecting to the input of panel.

Quoted 79 lines

Quoted 11 lines

I do not see why do you have such issue, could you describe it more.

Quoted 9 lines

Quoted 4 lines

I've at least seen one device (toshiba,tc358767 bridge) that can actually take either DPI as input or DSI based on how you configure it. There are 2 input ports here, port #0 for DSI and port #1 for DPI. Would it have made sense here to have a single port and 2 endpoints here too?

Nope, I guess having a port for DPI input, one port for DSI input makes quite a lot of sense. And then you can have the input-specifics living in the endpoints like dual links or so.

Quoted 8 lines

Quoted 5 lines

Quoted 5 lines

Quoted 3 lines

While the following argument might not be 100% valid from a dt-purity standpoint implementing this might get hairy for _any_ operating system, as you will need each panel/bridge to tell what the ports are used for.

Yeah.

Each panel/bridge should know what his ports are used for, but to who and why you want to communicate it?

I'll do that below :-)

Quoted 30 lines

Quoted 7 lines

Quoted 7 lines

Quoted 6 lines

I.e. in my endpoint based mapping, right now I have this nice and generic
WIP function to parse the of_graph and get the master+slave nodes:

https://github.com/mmind/linux-rockchip/blob/tmp/rk3399-gru-bob-scarlet/
drivers/gpu/drm/rockchip/dw-mipi-dsi-rockchip.c#L697 [0]

Quoted 22 lines

Quoted 3 lines

I'd tried out something locally before posting this patch, I don't have the code for it, but I can describe the steps I took. This code expects the panel/bridge to have 2 input ports.

Which again would be very much panel-specific as you cannot assume to much about the ports.

Quoted 15 lines

1. DSI0 host driver looks up its output port, gets the remote endpoint, and get this endpoint's parent (using of_graph_get_remote_port_parent()). Keeps the parent device node in a temp variable.

2. DSI1 host driver does the same thing.

3. DSI0 and DSI1 check whether their outputs are connected to the same device. If so, they're in dual DSI mode. If not, they are operating independently.

The positive of this approach is that we don't need to make any assumptions about the panel/bridge's port numbers, which is great. The negative is that our DSI controller instances now need to query each other, which can be messy, but not too hard to implement.

Quoted 4 lines

I think detection of dual dsi and whole configuration thing of dual-dsi should be done using in-kernel frameworks (maybe new ones :) ). Could you describe necessary configuration steps to be performed, what info should be passed from who to who, etc. It would be easier to discuss.

The base setup is the rk3399 soc, which has 2 completely separate dw-mipi-dsi controllers.

Gru-Scarlet (a ChromeOS tablet device) uses a high-res dual-DSI display (2500x1500 pixels or so) which is driven by both controllers together to achieve the required number of lanes to it.

So you end up with one (specific) controller being the master and the other being the slave, deferring to the master for most everything. As the controllers are somehow build to work together in that case via special soc-specific settings.

See [1] for current WIP code, I lifted out of the chromeos kernel and modified to not have that special rockchip,dual-channel =<&...>; property. That patch was on the lists already last year I think.

So each dw-mipi-dsi instance needs to at least know that it is in a master-slave-relationship and right now also needs access to the other driver instance. Doing the master->slave access should be fine in that special case, as the IP is the same type.

Quoted 2 lines

Also the crtc<->dsi interaction is quite a bit of handling-different between one crtc talking to 2 DSIs or 2 crtc talking to the 2 DSIs separately.

That's probably the bigger key: to treat them as completely separate ports means that you get separate CRTCs, IIUC (I'll admit, I'm still a bit rusty on navigating some DRM concepts).

Am Donnerstag, 7. Juni 2018, 23:10:20 CEST schrieb Brian Norris:

Quoted 3 lines

Hi,

I only have a little bit to add, but Heiko did solicit my opinion.

yep ... and I realized that I am/was way to attached to my (working) endpoint-based thingy to really appreciate the other arguments ;-)

And your DSI-related writings below, did provide some new thought- directions for me, so thanks for that.

Quoted 106 lines · Heiko Stuebner

Quoted 57 lines

Am Donnerstag, 7. Juni 2018, 12:39:03 CEST schrieb Andrzej Hajda:

Quoted 56 lines · Heiko Stuebner

Quoted 55 lines

Am Mittwoch, 6. Juni 2018, 18:07:46 CEST schrieb Archit Taneja:

Quoted 54 lines · Heiko Stübner

Quoted 40 lines

Am Mittwoch, 6. Juni 2018, 12:21:16 CEST schrieb Archit Taneja:

Quoted 39 lines · Heiko Stübner

Quoted 20 lines

Am Mittwoch, 6. Juni 2018, 07:59:29 CEST schrieb Archit Taneja:

Quoted 11 lines · Heiko Stuebner

Quoted 10 lines

Am Donnerstag, 18. Januar 2018, 05:53:55 CEST schrieb Archit Taneja:

Quoted 6 lines

Add binding info for peripherals that support dual-channel DSI. Add corresponding optional bindings for DSI host controllers that may be configured in this mode. Add an example of an I2C controlled device operating in dual-channel DSI mode.

Signed-off-by Archit Taneja <[email protected]>

Looks like a great solution for that problem, so

Reviewed-by Heiko Stuebner <[email protected]>

Quoted 8 lines

Quoted 5 lines

As I'm looking into that for my rk3399-scarlet device right now and couldn't find this patchset in the kernel yet, is it planned to merge or refresh these binding changes or were problems encountered.

At least an Ack/Review from Rob seems to be missing.

I forgot about these patches. Rob had reviewed the first one in the set the second one still needed an Ack. I'll post a v3 that adds the Reviewed-bys and fixes a small typo.

Quoted 7 lines

very nice ... because it looks like yesterday I managed to make the Rockchip dsi work in dual mode following this.

But one question came up, do you really want two input ports on the panel
side? I.e. hardware-wise, I guess the panel will have one 8-lane or so input thatonly gets split up on the soc side onto 2 dsi controllers?

I think all dual DSI panels actually have 2 DSI controllers/parsers within them, one on each port. The MIPI DSI spec doesn't support 8 lanes. Also, the pixels coming out of the host are distributed among the lanes differently than what would have been the case with a 'theoretical' 8 lane receiver.

Other than that, some dual DSI panels only accept DSI commands on the 'master' port, where as others expect the same command to be sent across
both the ports.

Quoted 6 lines

Quoted 4 lines

Therefore, I think it's better to represent dual DSI panels having 2 DSI input ports.

Your DT looks good to me.

Hmm, that doesn't match up then ;-) ... as my dt uses 2 endpoints in one port for the dsi-links.

Sorry, I didn't notice you'd created two endpoints within a single port.

I don't think I'm particular about 2 ports vs 1 port with 2 endpoints. They both seem okay to me as long as we follow it consistently. I'm myself not 100% sure of how to figure where one should prefer endpoints over ports. Maybe someone more familiar with the of graph bindings could comment here.

Quoted 4 lines

Quoted 3 lines

Port should describe physical port, endpoint are used to describe multiple links to the same port. If the panel has two physical DSI interfaces it should use two ports.

That again leaves a bit of space for interpretation ;-) .

Does it really? DSI maxes out at 4 lanes, so more than 4 lanes is clearly not a single standard port. Additionally, the code we're currently using in Chrome OS actually sends the same commands to each DSI separately (I'm not sure if it's actually required by the panels were using), so again, they are to some extent logically independent.

Quoted 3 lines

A dual-dsi display is probably pretty useless with only one controller connected to it, as its 4 lanes cannot satisfy the required 8 lanes of the panel.

Probably, although technically you can usually display on half the screen, if you have only have 1 working DSI.

Quoted 5 lines

See [0] for current WIP panel addition.
[Was already on the lists but needs cleanup]

Quoted 22 lines

[0] https://github.com/mmind/linux-rockchip/commit/459bc1a1599377c2c5408724e82619a4602a953d#diff-5d6d6ddab4fd282102d23d2c02e835f8R381

Quoted 19 lines

Quoted 8 lines

Quoted 8 lines

Quoted 8 lines

The issue I see with using ports and not endpoints for dual-dsi links is with distinguishing between input and output ports.

For a panel that's easy, as you every port will be an input port and if you have 2, it's supposed dual-dsi. But for example I guess there might exist some dual-dsi-to-something bridges, where you would end up with say 3 (or even more) ports ... two dual-dsi inputs and 1 or more outputs.

Quoted 5 lines

Quoted 3 lines

Okay, I get your point here. Although, even if the remote device had exactly 2 ports. Is it safe to assume that port #0 is an input port and port #1 is an output port? Is that the norm?

I don't think that anything like that is specified anywhere, so you cannot assume anything about what a port contains.

Ports are defined per device-node and port number assignment should be described in particular device bindings. So device driver knows well which functions should be assigned to which port. But it is private matter of the device/device driver. Different device bindings can assign different number for input and output ports, but since ports are private thing for devices it should not be a problem.

Agreed. And if Heiko needs a specific example close to home: all the Rockchip display bindings do this similarly; for example display/rockchip/analogix_dp-rockchip.txt describes exactly which port numbers and enpoints mean what:

Quoted 4 lines

- ports: there are 2 port nodes with endpoint definitions as defined in Documentation/devicetree/bindings/media/video-interfaces.txt. Port 0: contained 2 endpoints, connecting to the output of vop. Port 1: contained 1 endpoint, connecting to the input of panel.

Agreed here. My confusion stems from the issue that with the way the code is right now, the dsi controller would need to know how the ports of the panel are structured, to find its way to the 2nd controller.

But then again, maybe the implementation just needs to change ;-) .

Quoted 73 lines

Quoted 73 lines

Quoted 10 lines

I do not see why do you have such issue, could you describe it more.

Quoted 8 lines

Quoted 4 lines

I've at least seen one device (toshiba,tc358767 bridge) that can actually take either DPI as input or DSI based on how you configure it. There are 2 input ports here, port #0 for DSI and port #1 for DPI. Would it have made sense here to have a single port and 2 endpoints here too?

Nope, I guess having a port for DPI input, one port for DSI input makes quite a lot of sense. And then you can have the input-specifics living in the endpoints like dual links or so.

Quoted 6 lines

Quoted 4 lines

Quoted 4 lines

Quoted 3 lines

While the following argument might not be 100% valid from a dt-purity standpoint implementing this might get hairy for _any_ operating system, as you will need each panel/bridge to tell what the ports are used for.

Yeah.

Each panel/bridge should know what his ports are used for, but to who and why you want to communicate it?

I'll do that below :-)

Quoted 23 lines

Quoted 23 lines

Quoted 9 lines

Quoted 6 lines

I.e. in my endpoint based mapping, right now I have this nice and generic
WIP function to parse the of_graph and get the master+slave nodes:

https://github.com/mmind/linux-rockchip/blob/tmp/rk3399-gru-bob-scarlet/
drivers/gpu/drm/rockchip/dw-mipi-dsi-rockchip.c#L697 [0]

I'd tried out something locally before posting this patch, I don't have the code for it, but I can describe the steps I took. This code expects the panel/bridge to have 2 input ports.

Which again would be very much panel-specific as you cannot assume to much about the ports.

Quoted 11 lines

1. DSI0 host driver looks up its output port, gets the remote endpoint, and get this endpoint's parent (using of_graph_get_remote_port_parent()). Keeps the parent device node in a temp variable.

2. DSI1 host driver does the same thing.

3. DSI0 and DSI1 check whether their outputs are connected to the same device. If so, they're in dual DSI mode. If not, they are operating independently.

Quoted 8 lines

Quoted 4 lines

Quoted 4 lines

The positive of this approach is that we don't need to make any assumptions about the panel/bridge's port numbers, which is great. The negative is that our DSI controller instances now need to query each other, which can be messy, but not too hard to implement.

I think detection of dual dsi and whole configuration thing of dual-dsi should be done using in-kernel frameworks (maybe new ones :) ). Could you describe necessary configuration steps to be performed, what info should be passed from who to who, etc. It would be easier to discuss.

The base setup is the rk3399 soc, which has 2 completely separate dw-mipi-dsi controllers.

Gru-Scarlet (a ChromeOS tablet device) uses a high-res dual-DSI display (2500x1500 pixels or so) which is driven by both controllers together to achieve the required number of lanes to it.

So you end up with one (specific) controller being the master and the other being the slave, deferring to the master for most everything. As the controllers are somehow build to work together in that case via special soc-specific settings.

See [1] for current WIP code, I lifted out of the chromeos kernel and modified to not have that special rockchip,dual-channel =<&...>; property. That patch was on the lists already last year I think.

So each dw-mipi-dsi instance needs to at least know that it is in a master-slave-relationship and right now also needs access to the other driver instance. Doing the master->slave access should be fine in that special case, as the IP is the same type.

Also the crtc<->dsi interaction is quite a bit of handling-different between one crtc talking to 2 DSIs or 2 crtc talking to the 2 DSIs separately.

Quoted 3 lines

That's probably the bigger key: to treat them as completely separate ports means that you get separate CRTCs, IIUC (I'll admit, I'm still a bit rusty on navigating some DRM concepts).

One thing I realized with your mention of DSI maxing out at 4 lanes is, that this makes it easy to detect the existence of a dual-dsi situation ... simply when the device reports 8 lanes.
(via of_find_mipi_dsi_device_by_node presumably)

As a dual-dsi situation requires a clock-master property in one, both master and slave also would be able to determine their master or slave status, thus the global dual-dsi config could be done by the master (GRF stuff in the Rockchip case)

The only thing that makes my head explode now, is how to make the slave actually react to settings sent to the master in a sane way.

But that is a drm-specific implementation-detail, I guess ;-) . And hopefully someone might have a great idea how this could be done better than in my current implementation.

Heiko

Quoted 9 lines

Am Donnerstag, 7. Juni 2018, 23:10:20 CEST schrieb Brian Norris:

Quoted 3 lines

Hi,

I only have a little bit to add, but Heiko did solicit my opinion.

yep ... and I realized that I am/was way to attached to my (working) endpoint-based thingy to really appreciate the other arguments ;-)

And your DSI-related writings below, did provide some new thought- directions for me, so thanks for that.

Quoted 8 lines

Quoted 8 lines · Heiko Stuebner

Quoted 4 lines

Am Donnerstag, 7. Juni 2018, 12:39:03 CEST schrieb Andrzej Hajda:

Quoted 1 line · Heiko Stuebner

Also the crtc<->dsi interaction is quite a bit of handling-different between one crtc talking to 2 DSIs or 2 crtc talking to the 2 DSIs separately.

That's probably the bigger key: to treat them as completely separate ports means that you get separate CRTCs, IIUC (I'll admit, I'm still a bit rusty on navigating some DRM concepts).

Quoted 4 lines

One thing I realized with your mention of DSI maxing out at 4 lanes is, that this makes it easy to detect the existence of a dual-dsi situation ... simply when the device reports 8 lanes.
(via of_find_mipi_dsi_device_by_node presumably)

Hmm, device has two DSI interfaces, on each it has 4 lanes, so it report 4 lanes.

There is different problem with current implementation of panel lookup code - drm_panel is identified by device_node of physical device, as a result there can be only one panel per device.
I think proper identification should be by device_node of OF graph port, or by pair: device_node of physical device and port number (practically it is the same).
I think fixing it should not be a big deal.

Quoted 6 lines

As a dual-dsi situation requires a clock-master property in one, both master and slave also would be able to determine their master or slave status, thus the global dual-dsi config could be done by the master (GRF stuff in the Rockchip case)

Quoted 3 lines

The only thing that makes my head explode now, is how to make the slave actually react to settings sent to the master in a sane way.

Wouldn't be enough if the panel passes different bus info on DSI0 and DSI1?

Regards
Andrzej

Quoted 12 lines

But that is a drm-specific implementation-detail, I guess ;-) . And hopefully someone might have a great idea how this could be done better than in my current implementation.

Heiko

Add a section that describes dt-bindings for peripherals that support MIPI DSI, but have a different bus as the primary control bus, or no control bus at all. Add an example for a peripheral with a non-DSI control bus.

Reviewed-by Rob Herring <[email protected]>

Reviewed-by Boris Brezillon <[email protected]>

Reviewed-by Philippe Cornu <[email protected]>

Reviewed-by Sean Paul <[email protected]>

Signed-off-by Archit Taneja <[email protected]>

---
  .../devicetree/bindings/display/mipi-dsi-bus.txt   | 71 +++++++++++++++++++---
  1 file changed, 64 insertions(+), 7 deletions(-)

Documentation/devicetree/bindings/display/mipi-dsi-bus.txt +54 -7

index 973c27273772..b7c5bf47403c 100644
@@ -16,7 +16,7 @@ The following assumes that only a single peripheral is connected to a DSI
  host. Experience shows that this is true for the large majority of setups.
 
  DSI host
---------
+========
 
  In addition to the standard properties and those defined by the parent bus of
  a DSI host, the following properties apply to a node representing a DSI host.

@@ -30,11 +30,16 @@ Required properties:
    different value here. See below.
 
  DSI peripheral
---------------
+==============
 
-Peripherals are represented as child nodes of the DSI host's node. Properties
-described here apply to all DSI peripherals, but individual bindings may want
-to define additional, device-specific properties.
+Peripherals with DSI as control bus, or no control bus
+------------------------------------------------------
+
+Peripherals with the DSI bus as the primary control bus, or peripherals with
+no control bus but use the DSI bus to transmit pixel data are represented
+as child nodes of the DSI host's node. Properties described here apply to all
+DSI peripherals, but individual bindings may want to define additional,
+device-specific properties.
 
  Required properties:
  - reg: The virtual channel number of a DSI peripheral. Must be in the range

@@ -49,9 +54,25 @@ case two alternative representations can be chosen:
    property is the number of the first virtual channel and the second cell is
    the number of consecutive virtual channels.
 
-Example
--------
+Peripherals with a different control bus
+----------------------------------------
+
+There are peripherals that have I2C/SPI (or some other non-DSI bus) as the
+primary control bus, but are also connected to a DSI bus (mostly for the data
+path). Connections between such peripherals and a DSI host can be represented
+using the graph bindings [1], [2].
+
+[1] Documentation/devicetree/bindings/graph.txt
+[2] Documentation/devicetree/bindings/media/video-interfaces.txt
 
+Examples
+========
+- (1), (2) and (3) are examples of a DSI host and peripheral on the DSI bus
+  with different virtual channel configurations.
+- (4) is an example of a peripheral on a I2C control bus connected to a
+  DSI host using of-graph bindings.
+
+1)
  	dsi-host {
...

@@ -67,6 +88,7 @@ Example
...
  	};
 
+2)
  	dsi-host {
...

@@ -82,6 +104,7 @@ Example
...
  	};
 
+3)
  	dsi-host {
...

@@ -96,3 +119,37 @@ Example
 
...
  	};
+
+4)
+	i2c-host {
+		...
+
+		dsi-bridge@35 {
+			compatible = "...";
+			reg = <0x35>;
+
+			ports {
+				...
+
+				port {
+					bridge_mipi_in: endpoint {
+						remote-endpoint = <&host_mipi_out>;
+					};
+				};
+			};
+		};
+	};
+
+	dsi-host {
+		...
+

Add binding info for peripherals that support dual-channel DSI. Add corresponding optional bindings for DSI host controllers that may be configured in this mode. Add an example of an I2C controlled device operating in dual-channel DSI mode.

Reviewed-by Philippe Cornu <[email protected]>

Reviewed-by Sean Paul <[email protected]>

Reviewed-by Heiko Stuebner <[email protected]>

Signed-off-by Archit Taneja <[email protected]>

---
 .../devicetree/bindings/display/mipi-dsi-bus.txt   | 80 ++++++++++++++++++++++
 1 file changed, 80 insertions(+)

Quoted 5 lines · Archit Taneja

Add binding info for peripherals that support dual-channel DSI. Add corresponding optional bindings for DSI host controllers that may be configured in this mode. Add an example of an I2C controlled device operating in dual-channel DSI mode.

Quoted 7 lines

Reviewed-by Philippe Cornu <[email protected]>

Reviewed-by Sean Paul <[email protected]>

Reviewed-by Heiko Stuebner <[email protected]>

Signed-off-by Archit Taneja <[email protected]>

---
  .../devicetree/bindings/display/mipi-dsi-bus.txt   | 80 ++++++++++++++++++++++
  1 file changed, 80 insertions(+)

Reviewed-by Rob Herring <[email protected]>