Skip to contentSkip to conversation

leds: add driver for SPI driven WorldSemi WS2812B RGB LEDs

PATCH In review Chuanhong Guo 4 participants 17 messages Dec 12, 2022 – Jan 20, 2023

Subsystems (3)
lore
Share & tools
lore
#0 lore raw

v4 cover leds: add driver for SPI driven WorldSemi WS2812B RGB LEDs

This patch adds support for driving a chain of WS2812B LED chips using SPI bus.

WorldSemi WS2812B is a individually addressable LED chip that can be chained together and controlled individually using a single wire. The chip recognize a long pulse as a bit of 1 and a short pulse as a bit of 0. Host sends a continuous stream of 24-bits color values, each LED chip takes the first 3 byte it receives as its color value and passes the leftover bytes to the next LED on the chain.

This driver simulates this protocol using SPI bus by sending a long pulse as 3'b110 and a short pulse as 3'b100. The SPI frequency needs to be 2.105MHz~2.85MHz for the timing to be correct and the controller needs to transfer all the bytes continuously.

Changes since v1:
various dt binding fixes
add support for default-brightness

Changes since v2:
more dt binding fixes
drop default-brightness and default-intensity

Changes since v3:
1. add more comments
2. rename reg to cascade
3. redo some line breaking
4. move duplicated pointer calculation into ws2812b_set_byte
5. reword error message
6. get ws2812b_priv from led cdev->dev->parent

Chuanhong Guo (3):
dt-bindings: vendor-prefixes: add an entry for WorldSemi
dt-bindings: leds: add worldsemi,ws2812b
leds: add driver for SPI driven WorldSemi WS2812B RGB LEDs

 .../bindings/leds/worldsemi,ws2812b.yaml      | 116 +++++++++
 .../devicetree/bindings/vendor-prefixes.yaml  |   2 +
 drivers/leds/rgb/Kconfig                      |  11 +
 drivers/leds/rgb/Makefile                     |   1 +
 drivers/leds/rgb/leds-ws2812b.c               | 231 ++++++++++++++++++
 5 files changed, 361 insertions(+)
 create mode 100644 Documentation/devicetree/bindings/leds/worldsemi,ws2812b.yaml
 create mode 100644 drivers/leds/rgb/leds-ws2812b.c
Signature · 1 line
2.38.1
#1 lore raw

PATCH v4 1/3 dt-bindings: vendor-prefixes: add an entry for WorldSemi

Add vendor prefix for WorldSemi that makes WS2812B individually-addressable RGB LEDs.

---
Change since v1:
reword commit message
Change since v2: none

 Documentation/devicetree/bindings/vendor-prefixes.yaml | 2 ++
 1 file changed, 2 insertions(+)
Documentation/devicetree/bindings/vendor-prefixes.yaml +2 -0
index 10c178d97b02..32274d894664 100644
@@ -1462,6 +1462,8 @@ patternProperties:
     description: Wondermedia Technologies, Inc.
   "^wobo,.*":
     description: Wobo
+  "^worldsemi,.*":
+    description: WorldSemi Co., Limited
   "^wanchanglong,.*":
     description: Wanchanglong Electronics Technology(SHENZHEN)Co.,Ltd.
   "^x-powers,.*":
Signature · 1 line
2.38.1
#2 lore raw

PATCH v4 2/3 dt-bindings: leds: add worldsemi,ws2812b

In reply to Chuanhong Guo

Add dt binding schema for WorldSemi WS2812B driven using SPI bus.

---
Changes since v1:
remove linux driver reference from description remove some obvious descriptions
fix unit address regex in multi-led property drop various minItems
add maxItems = 1 to reg
fix node names and property orders in binding example drop -spi from compatible string
add default-brightness

Change since v2:
drop "this patch" from commit message
rename leds to led-controller
drop default-brightness and default-intensity

Change since v3:
reword commit title

 .../bindings/leds/worldsemi,ws2812b.yaml      | 116 ++++++++++++++++++
 1 file changed, 116 insertions(+)
 create mode 100644 Documentation/devicetree/bindings/leds/worldsemi,ws2812b.yaml
Documentation/devicetree/bindings/leds/worldsemi,ws2812b.yaml new file +116 -0
new file mode 100644
index 000000000000..548c05ac3d31
@@ -0,0 +1,116 @@
+# SPDX-License-Identifier: (GPL-2.0-only OR BSD-2-Clause)
+%YAML 1.2
+---
+$id: http://devicetree.org/schemas/leds/worldsemi,ws2812b.yaml#
+$schema: http://devicetree.org/meta-schemas/core.yaml#
+
+title: WS2812B LEDs driven using SPI
+
+maintainers:
+  - Chuanhong Guo <[email protected]>
+
+description: |
+  WorldSemi WS2812B is a individually addressable LED chip that can be chained
+  together and controlled individually using a single wire.
+  This binding describes a chain of WS2812B LEDs connected to the SPI MOSI pin.
+  Typical setups includes connecting the data pin of the LED chain to MOSI as
+  the only device or using CS and MOSI with a tri-state voltage-level shifter
+  for the data pin.
+  The SPI frequency needs to be 2.105MHz~2.85MHz for the timing to be correct
+  and the controller needs to send all the bytes continuously.
+
+allOf:
+  - $ref: /schemas/spi/spi-peripheral-props.yaml#
+
+properties:
+  compatible:
+    const: worldsemi,ws2812b
+
+  reg:
+    maxItems: 1
+
+  spi-max-frequency:
+    minimum: 2105000
+    maximum: 2850000
+
+  "#address-cells":
+    const: 1
+
+  "#size-cells":
+    const: 0
+
+patternProperties:
+  "^multi-led@[0-9a-f]+$":
+    type: object
+    $ref: leds-class-multicolor.yaml#
+    unevaluatedProperties: false
+
+    properties:
+      color-index:
+        description: |
+          A 3-item array specifying color of each components in this LED. It
+          should be one of the LED_COLOR_ID_* prefixed definitions from the
+          header include/dt-bindings/leds/common.h. Defaults to
+          <LED_COLOR_ID_GREEN LED_COLOR_ID_RED LED_COLOR_ID_BLUE>
+          if unspecified.
+        $ref: /schemas/types.yaml#/definitions/uint32-array
+        maxItems: 3
+
+      reg:
+        description: |
+          Which LED this node represents. The reg of the first LED on the chain
+          is 0.
+        maxItems: 1
+
+    required:
+      - reg
+      - color
+      - function
+
+required:
+  - compatible
+
+additionalProperties: false
+
+examples:
+  - |
+    #include <dt-bindings/leds/common.h>
+    spi {
+        #address-cells = <1>;
+        #size-cells = <0>;
+
+        led-controller@0 {
+            compatible = "worldsemi,ws2812b";
+            reg = <0>;
+            #address-cells = <1>;
+            #size-cells = <0>;
+            spi-max-frequency = <2850000>;
+            multi-led@0 {
+                reg = <0>;
+                color-index = <LED_COLOR_ID_RED LED_COLOR_ID_GREEN LED_COLOR_ID_BLUE>;
+                color = <LED_COLOR_ID_RGB>;
+                function = LED_FUNCTION_STATUS;
+                function-enumerator = <0>;
+            };
+            multi-led@1 {
+                reg = <1>;
+                color = <LED_COLOR_ID_RGB>;
+                function = LED_FUNCTION_STATUS;
+                function-enumerator = <1>;
+            };
+            multi-led@2 {
+                reg = <2>;
+                color = <LED_COLOR_ID_RGB>;
+                function = LED_FUNCTION_STATUS;
+                function-enumerator = <2>;
+            };
+            multi-led@3 {
+                reg = <3>;
+                color = <LED_COLOR_ID_RGB>;
+                function = LED_FUNCTION_STATUS;
+                function-enumerator = <3>;
+            };
+        };
+    };
+
+...
Signature · 1 line
2.38.1
#3 lore raw

Re: [PATCH v4 2/3] dt-bindings: leds: add worldsemi,ws2812b

Quoted 80 lines · Chuanhong Guo

Add dt binding schema for WorldSemi WS2812B driven using SPI bus.

---
Changes since v1:
remove linux driver reference from description remove some obvious descriptions
fix unit address regex in multi-led property drop various minItems
add maxItems = 1 to reg
fix node names and property orders in binding example drop -spi from compatible string
add default-brightness

Change since v2:
drop "this patch" from commit message
rename leds to led-controller
drop default-brightness and default-intensity

Change since v3:
reword commit title

 .../bindings/leds/worldsemi,ws2812b.yaml      | 116 ++++++++++++++++++
 1 file changed, 116 insertions(+)
 create mode 100644 Documentation/devicetree/bindings/leds/worldsemi,ws2812b.yaml
Documentation/devicetree/bindings/leds/worldsemi,ws2812b.yaml new file +47 -0
new file mode 100644
index 000000000000..548c05ac3d31
@@ -0,0 +1,116 @@
+# SPDX-License-Identifier: (GPL-2.0-only OR BSD-2-Clause)
+%YAML 1.2
+---
+$id: http://devicetree.org/schemas/leds/worldsemi,ws2812b.yaml#
+$schema: http://devicetree.org/meta-schemas/core.yaml#
+
+title: WS2812B LEDs driven using SPI
+
+maintainers:
+  - Chuanhong Guo <[email protected]>
+
+description: |
+  WorldSemi WS2812B is a individually addressable LED chip that can be chained
+  together and controlled individually using a single wire.
+  This binding describes a chain of WS2812B LEDs connected to the SPI MOSI pin.
+  Typical setups includes connecting the data pin of the LED chain to MOSI as
+  the only device or using CS and MOSI with a tri-state voltage-level shifter
+  for the data pin.
+  The SPI frequency needs to be 2.105MHz~2.85MHz for the timing to be correct
+  and the controller needs to send all the bytes continuously.
+
+allOf:
+  - $ref: /schemas/spi/spi-peripheral-props.yaml#
+
+properties:
+  compatible:
+    const: worldsemi,ws2812b
+
+  reg:
+    maxItems: 1
+
+  spi-max-frequency:
+    minimum: 2105000
+    maximum: 2850000
+
+  "#address-cells":
+    const: 1
+
+  "#size-cells":
+    const: 0
+
+patternProperties:
+  "^multi-led@[0-9a-f]+$":
+    type: object
+    $ref: leds-class-multicolor.yaml#
+    unevaluatedProperties: false
+

Documentation/devicetree/bindings/leds/worldsemi,ws2812b.yaml line 57

Quoted 10 lines
Documentation/devicetree/bindings/leds/worldsemi,ws2812b.yaml new file +10 -0
+    properties:
+      color-index:
+        description: |
+          A 3-item array specifying color of each components in this LED. It
+          should be one of the LED_COLOR_ID_* prefixed definitions from the
+          header include/dt-bindings/leds/common.h. Defaults to
+          <LED_COLOR_ID_GREEN LED_COLOR_ID_RED LED_COLOR_ID_BLUE>
+          if unspecified.
+        $ref: /schemas/types.yaml#/definitions/uint32-array
+        maxItems: 3

In general I am fine with it, although there is still question for adding more multi-color defines in binding headers to replace this property - GRB/RBG/GBR and even more for RGBW.

Pavel, Lee, any thoughts from your side?

Best regards,
Krzysztof

maintainer
#4 lore raw

Re: [PATCH v4 2/3] dt-bindings: leds: add worldsemi,ws2812b

Quoted 96 lines · Krzysztof Kozlowski
Quoted 80 lines · Chuanhong Guo

Add dt binding schema for WorldSemi WS2812B driven using SPI bus.

---
Changes since v1:
remove linux driver reference from description remove some obvious descriptions
fix unit address regex in multi-led property drop various minItems
add maxItems = 1 to reg
fix node names and property orders in binding example drop -spi from compatible string
add default-brightness

Change since v2:
drop "this patch" from commit message
rename leds to led-controller
drop default-brightness and default-intensity

Change since v3:
reword commit title

 .../bindings/leds/worldsemi,ws2812b.yaml      | 116 ++++++++++++++++++
 1 file changed, 116 insertions(+)
 create mode 100644 Documentation/devicetree/bindings/leds/worldsemi,ws2812b.yaml
Documentation/devicetree/bindings/leds/worldsemi,ws2812b.yaml new file +47 -0
new file mode 100644
index 000000000000..548c05ac3d31
@@ -0,0 +1,116 @@
+# SPDX-License-Identifier: (GPL-2.0-only OR BSD-2-Clause)
+%YAML 1.2
+---
+$id: http://devicetree.org/schemas/leds/worldsemi,ws2812b.yaml#
+$schema: http://devicetree.org/meta-schemas/core.yaml#
+
+title: WS2812B LEDs driven using SPI
+
+maintainers:
+  - Chuanhong Guo <[email protected]>
+
+description: |
+  WorldSemi WS2812B is a individually addressable LED chip that can be chained
+  together and controlled individually using a single wire.
+  This binding describes a chain of WS2812B LEDs connected to the SPI MOSI pin.
+  Typical setups includes connecting the data pin of the LED chain to MOSI as
+  the only device or using CS and MOSI with a tri-state voltage-level shifter
+  for the data pin.
+  The SPI frequency needs to be 2.105MHz~2.85MHz for the timing to be correct
+  and the controller needs to send all the bytes continuously.
+
+allOf:
+  - $ref: /schemas/spi/spi-peripheral-props.yaml#
+
+properties:
+  compatible:
+    const: worldsemi,ws2812b
+
+  reg:
+    maxItems: 1
+
+  spi-max-frequency:
+    minimum: 2105000
+    maximum: 2850000
+
+  "#address-cells":
+    const: 1
+
+  "#size-cells":
+    const: 0
+
+patternProperties:
+  "^multi-led@[0-9a-f]+$":
+    type: object
+    $ref: leds-class-multicolor.yaml#
+    unevaluatedProperties: false
+

Documentation/devicetree/bindings/leds/worldsemi,ws2812b.yaml line 57

Quoted 10 lines
Documentation/devicetree/bindings/leds/worldsemi,ws2812b.yaml new file +10 -0
+    properties:
+      color-index:
+        description: |
+          A 3-item array specifying color of each components in this LED. It
+          should be one of the LED_COLOR_ID_* prefixed definitions from the
+          header include/dt-bindings/leds/common.h. Defaults to
+          <LED_COLOR_ID_GREEN LED_COLOR_ID_RED LED_COLOR_ID_BLUE>
+          if unspecified.
+        $ref: /schemas/types.yaml#/definitions/uint32-array
+        maxItems: 3

In general I am fine with it, although there is still question for adding more multi-color defines in binding headers to replace this property - GRB/RBG/GBR and even more for RGBW.

Quoted 1 line

Pavel, Lee, any thoughts from your side?

Nothing from me yet.

If Pavel doesn't respond soon, I'll get around to doing a full sweep after the merge-window has closed. Actually, most likely the new year now.

Signature · 1 line
Lee Jones [李琼斯]
#5 lore raw

Re: [PATCH v4 2/3] dt-bindings: leds: add worldsemi,ws2812b

In reply to Krzysztof Kozlowski

Hi!

Quoted 95 lines
Quoted 80 lines

Add dt binding schema for WorldSemi WS2812B driven using SPI bus.

---
Changes since v1:
remove linux driver reference from description remove some obvious descriptions
fix unit address regex in multi-led property drop various minItems
add maxItems = 1 to reg
fix node names and property orders in binding example drop -spi from compatible string
add default-brightness

Change since v2:
drop "this patch" from commit message
rename leds to led-controller
drop default-brightness and default-intensity

Change since v3:
reword commit title

 .../bindings/leds/worldsemi,ws2812b.yaml      | 116 ++++++++++++++++++
 1 file changed, 116 insertions(+)
 create mode 100644 Documentation/devicetree/bindings/leds/worldsemi,ws2812b.yaml
Documentation/devicetree/bindings/leds/worldsemi,ws2812b.yaml new file +47 -0
new file mode 100644
index 000000000000..548c05ac3d31
@@ -0,0 +1,116 @@
+# SPDX-License-Identifier: (GPL-2.0-only OR BSD-2-Clause)
+%YAML 1.2
+---
+$id: http://devicetree.org/schemas/leds/worldsemi,ws2812b.yaml#
+$schema: http://devicetree.org/meta-schemas/core.yaml#
+
+title: WS2812B LEDs driven using SPI
+
+maintainers:
+  - Chuanhong Guo <[email protected]>
+
+description: |
+  WorldSemi WS2812B is a individually addressable LED chip that can be chained
+  together and controlled individually using a single wire.
+  This binding describes a chain of WS2812B LEDs connected to the SPI MOSI pin.
+  Typical setups includes connecting the data pin of the LED chain to MOSI as
+  the only device or using CS and MOSI with a tri-state voltage-level shifter
+  for the data pin.
+  The SPI frequency needs to be 2.105MHz~2.85MHz for the timing to be correct
+  and the controller needs to send all the bytes continuously.
+
+allOf:
+  - $ref: /schemas/spi/spi-peripheral-props.yaml#
+
+properties:
+  compatible:
+    const: worldsemi,ws2812b
+
+  reg:
+    maxItems: 1
+
+  spi-max-frequency:
+    minimum: 2105000
+    maximum: 2850000
+
+  "#address-cells":
+    const: 1
+
+  "#size-cells":
+    const: 0
+
+patternProperties:
+  "^multi-led@[0-9a-f]+$":
+    type: object
+    $ref: leds-class-multicolor.yaml#
+    unevaluatedProperties: false
+

Documentation/devicetree/bindings/leds/worldsemi,ws2812b.yaml line 57

Quoted 10 lines
Documentation/devicetree/bindings/leds/worldsemi,ws2812b.yaml new file +10 -0
+    properties:
+      color-index:
+        description: |
+          A 3-item array specifying color of each components in this LED. It
+          should be one of the LED_COLOR_ID_* prefixed definitions from the
+          header include/dt-bindings/leds/common.h. Defaults to
+          <LED_COLOR_ID_GREEN LED_COLOR_ID_RED LED_COLOR_ID_BLUE>
+          if unspecified.
+        $ref: /schemas/types.yaml#/definitions/uint32-array
+        maxItems: 3

In general I am fine with it, although there is still question for adding more multi-color defines in binding headers to replace this property - GRB/RBG/GBR and even more for RGBW.

Quoted 1 line

Pavel, Lee, any thoughts from your side?

This really needs to mention the name this hardware is known as -- I believe it is NeoPixel.

Best regards,
Pavel

Signature · 1 line
People of Russia, stop Putin before his war on Ukraine escalates.
#6 lore raw

Re: [PATCH v4 2/3] dt-bindings: leds: add worldsemi,ws2812b

Hi!

Quoted 8 lines · Pavel Machek
Quoted 5 lines

In general I am fine with it, although there is still question for adding more multi-color defines in binding headers to replace this property - GRB/RBG/GBR and even more for RGBW.

Pavel, Lee, any thoughts from your side?

This really needs to mention the name this hardware is known as -- I believe it is NeoPixel.

I'll do so in the next version. I'm actually waiting for an answer to the comment from Krzysztof (whether I should change color-index) before sending v5 :)

Signature · 2 lines
Regards,
Chuanhong Guo
#7 lore raw

Re: [PATCH v4 2/3] dt-bindings: leds: add worldsemi,ws2812b

In reply to Pavel Machek
Quoted 97 lines · Pavel Machek

Hi!

Quoted 95 lines
Quoted 80 lines

Add dt binding schema for WorldSemi WS2812B driven using SPI bus.

---
Changes since v1:
remove linux driver reference from description remove some obvious descriptions
fix unit address regex in multi-led property drop various minItems
add maxItems = 1 to reg
fix node names and property orders in binding example drop -spi from compatible string
add default-brightness

Change since v2:
drop "this patch" from commit message
rename leds to led-controller
drop default-brightness and default-intensity

Change since v3:
reword commit title

 .../bindings/leds/worldsemi,ws2812b.yaml      | 116 ++++++++++++++++++
 1 file changed, 116 insertions(+)
 create mode 100644 Documentation/devicetree/bindings/leds/worldsemi,ws2812b.yaml
Documentation/devicetree/bindings/leds/worldsemi,ws2812b.yaml new file +47 -0
new file mode 100644
index 000000000000..548c05ac3d31
@@ -0,0 +1,116 @@
+# SPDX-License-Identifier: (GPL-2.0-only OR BSD-2-Clause)
+%YAML 1.2
+---
+$id: http://devicetree.org/schemas/leds/worldsemi,ws2812b.yaml#
+$schema: http://devicetree.org/meta-schemas/core.yaml#
+
+title: WS2812B LEDs driven using SPI
+
+maintainers:
+  - Chuanhong Guo <[email protected]>
+
+description: |
+  WorldSemi WS2812B is a individually addressable LED chip that can be chained
+  together and controlled individually using a single wire.
+  This binding describes a chain of WS2812B LEDs connected to the SPI MOSI pin.
+  Typical setups includes connecting the data pin of the LED chain to MOSI as
+  the only device or using CS and MOSI with a tri-state voltage-level shifter
+  for the data pin.
+  The SPI frequency needs to be 2.105MHz~2.85MHz for the timing to be correct
+  and the controller needs to send all the bytes continuously.
+
+allOf:
+  - $ref: /schemas/spi/spi-peripheral-props.yaml#
+
+properties:
+  compatible:
+    const: worldsemi,ws2812b
+
+  reg:
+    maxItems: 1
+
+  spi-max-frequency:
+    minimum: 2105000
+    maximum: 2850000
+
+  "#address-cells":
+    const: 1
+
+  "#size-cells":
+    const: 0
+
+patternProperties:
+  "^multi-led@[0-9a-f]+$":
+    type: object
+    $ref: leds-class-multicolor.yaml#
+    unevaluatedProperties: false
+

Documentation/devicetree/bindings/leds/worldsemi,ws2812b.yaml line 57

Quoted 10 lines
Documentation/devicetree/bindings/leds/worldsemi,ws2812b.yaml new file +10 -0
+    properties:
+      color-index:
+        description: |
+          A 3-item array specifying color of each components in this LED. It
+          should be one of the LED_COLOR_ID_* prefixed definitions from the
+          header include/dt-bindings/leds/common.h. Defaults to
+          <LED_COLOR_ID_GREEN LED_COLOR_ID_RED LED_COLOR_ID_BLUE>
+          if unspecified.
+        $ref: /schemas/types.yaml#/definitions/uint32-array
+        maxItems: 3

In general I am fine with it, although there is still question for adding more multi-color defines in binding headers to replace this property - GRB/RBG/GBR and even more for RGBW.

Quoted 4 lines
Quoted 1 line

Pavel, Lee, any thoughts from your side?

This really needs to mention the name this hardware is known as -- I believe it is NeoPixel.

We wait here for feedback on colors... The binding is re-implementing color, just because of combinations GRB/RBG/GBR, which could be achieved with new color defines.

Best regards,
Krzysztof

maintainer
#8 lore raw

Re: [PATCH v4 2/3] dt-bindings: leds: add worldsemi,ws2812b

Quoted 77 lines · Krzysztof Kozlowski
Quoted 76 lines · Pavel Machek

Hi!

Quoted 74 lines
Quoted 74 lines

Add dt binding schema for WorldSemi WS2812B driven using SPI bus.

---
Changes since v1:
remove linux driver reference from description remove some obvious descriptions
fix unit address regex in multi-led property drop various minItems
add maxItems = 1 to reg
fix node names and property orders in binding example drop -spi from compatible string
add default-brightness

Change since v2:
drop "this patch" from commit message
rename leds to led-controller
drop default-brightness and default-intensity

Change since v3:
reword commit title

 .../bindings/leds/worldsemi,ws2812b.yaml      | 116 ++++++++++++++++++
 1 file changed, 116 insertions(+)
 create mode 100644 Documentation/devicetree/bindings/leds/worldsemi,ws2812b.yaml
Documentation/devicetree/bindings/leds/worldsemi,ws2812b.yaml new file +41 -0
new file mode 100644
index 000000000000..548c05ac3d31
@@ -0,0 +1,116 @@
+# SPDX-License-Identifier: (GPL-2.0-only OR BSD-2-Clause)
+%YAML 1.2
+---
+$id: http://devicetree.org/schemas/leds/worldsemi,ws2812b.yaml#
+$schema: http://devicetree.org/meta-schemas/core.yaml#
+
+title: WS2812B LEDs driven using SPI
+
+maintainers:
+  - Chuanhong Guo <[email protected]>
+
+description: |
+  WorldSemi WS2812B is a individually addressable LED chip that can be chained
+  together and controlled individually using a single wire.
+  This binding describes a chain of WS2812B LEDs connected to the SPI MOSI pin.
+  Typical setups includes connecting the data pin of the LED chain to MOSI as
+  the only device or using CS and MOSI with a tri-state voltage-level shifter
+  for the data pin.
+  The SPI frequency needs to be 2.105MHz~2.85MHz for the timing to be correct
+  and the controller needs to send all the bytes continuously.
+
+allOf:
+  - $ref: /schemas/spi/spi-peripheral-props.yaml#
+
+properties:
+  compatible:
+    const: worldsemi,ws2812b
+
+  reg:
+    maxItems: 1
+
+  spi-max-frequency:
+    minimum: 2105000
+    maximum: 2850000
+
+  "#address-cells":
+    const: 1
+
+  "#size-cells":
+    const: 0
+
Quoted 8 lines
Quoted 8 lines
Quoted 8 lines

Documentation/devicetree/bindings/leds/worldsemi,ws2812b.yaml line 49

Quoted 8 lines
Documentation/devicetree/bindings/leds/worldsemi,ws2812b.yaml new file +8 -0
+patternProperties:
+  "^multi-led@[0-9a-f]+$":
+    type: object
+    $ref: leds-class-multicolor.yaml#
+    unevaluatedProperties: false
+
+    properties:
+      color-index:

Why "index"?

Isn't it just an array of colours rather than an index into something?

Quoted 2 lines
Quoted 2 lines
Quoted 2 lines

Documentation/devicetree/bindings/leds/worldsemi,ws2812b.yaml line 51

Quoted 2 lines
Documentation/devicetree/bindings/leds/worldsemi,ws2812b.yaml new file +2 -0
+        description: |
+          A 3-item array specifying color of each components in this LED. It

Why are you forcing this to 3?

Surely there are multi-colour LEDs containing more or less colours?

Quoted 2 lines
Quoted 2 lines
Quoted 2 lines

Documentation/devicetree/bindings/leds/worldsemi,ws2812b.yaml line 53

Quoted 2 lines
Documentation/devicetree/bindings/leds/worldsemi,ws2812b.yaml new file +2 -0
+          should be one of the LED_COLOR_ID_* prefixed definitions from the
+          header include/dt-bindings/leds/common.h. Defaults to

Isn't "include" a Linuxisum?

Quoted 9 lines
Quoted 9 lines
Quoted 9 lines

Documentation/devicetree/bindings/leds/worldsemi,ws2812b.yaml line 57

Quoted 4 lines
Documentation/devicetree/bindings/leds/worldsemi,ws2812b.yaml new file +4 -0
+          <LED_COLOR_ID_GREEN LED_COLOR_ID_RED LED_COLOR_ID_BLUE>
+          if unspecified.
+        $ref: /schemas/types.yaml#/definitions/uint32-array
+        maxItems: 3

In general I am fine with it, although there is still question for adding more multi-color defines in binding headers to replace this property - GRB/RBG/GBR and even more for RGBW.

Quoted 8 lines
Quoted 4 lines
Quoted 1 line

Pavel, Lee, any thoughts from your side?

This really needs to mention the name this hardware is known as -- I believe it is NeoPixel.

We wait here for feedback on colors... The binding is re-implementing color, just because of combinations GRB/RBG/GBR, which could be achieved with new color defines.

Sure, but where does that end?

How many permutations are there likely to be?

An unlimited array has more of a chance of standing the test of time.

Signature · 1 line
Lee Jones [李琼斯]
#9 lore raw

Re: [PATCH v4 2/3] dt-bindings: leds: add worldsemi,ws2812b

Quoted 90 lines · Lee Jones
Quoted 77 lines · Krzysztof Kozlowski
Quoted 76 lines · Pavel Machek

Hi!

Quoted 74 lines
Quoted 74 lines

Add dt binding schema for WorldSemi WS2812B driven using SPI bus.

---
Changes since v1:
remove linux driver reference from description remove some obvious descriptions
fix unit address regex in multi-led property drop various minItems
add maxItems = 1 to reg
fix node names and property orders in binding example drop -spi from compatible string
add default-brightness

Change since v2:
drop "this patch" from commit message
rename leds to led-controller
drop default-brightness and default-intensity

Change since v3:
reword commit title

 .../bindings/leds/worldsemi,ws2812b.yaml      | 116 ++++++++++++++++++
 1 file changed, 116 insertions(+)
 create mode 100644 Documentation/devicetree/bindings/leds/worldsemi,ws2812b.yaml
Documentation/devicetree/bindings/leds/worldsemi,ws2812b.yaml new file +41 -0
new file mode 100644
index 000000000000..548c05ac3d31
@@ -0,0 +1,116 @@
+# SPDX-License-Identifier: (GPL-2.0-only OR BSD-2-Clause)
+%YAML 1.2
+---
+$id: http://devicetree.org/schemas/leds/worldsemi,ws2812b.yaml#
+$schema: http://devicetree.org/meta-schemas/core.yaml#
+
+title: WS2812B LEDs driven using SPI
+
+maintainers:
+  - Chuanhong Guo <[email protected]>
+
+description: |
+  WorldSemi WS2812B is a individually addressable LED chip that can be chained
+  together and controlled individually using a single wire.
+  This binding describes a chain of WS2812B LEDs connected to the SPI MOSI pin.
+  Typical setups includes connecting the data pin of the LED chain to MOSI as
+  the only device or using CS and MOSI with a tri-state voltage-level shifter
+  for the data pin.
+  The SPI frequency needs to be 2.105MHz~2.85MHz for the timing to be correct
+  and the controller needs to send all the bytes continuously.
+
+allOf:
+  - $ref: /schemas/spi/spi-peripheral-props.yaml#
+
+properties:
+  compatible:
+    const: worldsemi,ws2812b
+
+  reg:
+    maxItems: 1
+
+  spi-max-frequency:
+    minimum: 2105000
+    maximum: 2850000
+
+  "#address-cells":
+    const: 1
+
+  "#size-cells":
+    const: 0
+
Quoted 8 lines
Quoted 8 lines
Quoted 8 lines

Documentation/devicetree/bindings/leds/worldsemi,ws2812b.yaml line 49

Quoted 8 lines
Documentation/devicetree/bindings/leds/worldsemi,ws2812b.yaml new file +8 -0
+patternProperties:
+  "^multi-led@[0-9a-f]+$":
+    type: object
+    $ref: leds-class-multicolor.yaml#
+    unevaluatedProperties: false
+
+    properties:
+      color-index:

Why "index"?

Quoted 1 line

Isn't it just an array of colours rather than an index into something?

Yeah.

Quoted 7 lines
Quoted 3 lines
Quoted 3 lines
Quoted 3 lines

Documentation/devicetree/bindings/leds/worldsemi,ws2812b.yaml line 51

Quoted 3 lines
Documentation/devicetree/bindings/leds/worldsemi,ws2812b.yaml new file +2 -0
+        description: |
+          A 3-item array specifying color of each components in this LED. It

Why are you forcing this to 3?

Surely there are multi-colour LEDs containing more or less colours?

For this device, because it has only tuples of three.

Quoted 5 lines
Quoted 3 lines
Quoted 3 lines
Quoted 3 lines

Documentation/devicetree/bindings/leds/worldsemi,ws2812b.yaml line 53

Quoted 3 lines
Documentation/devicetree/bindings/leds/worldsemi,ws2812b.yaml new file +2 -0
+          should be one of the LED_COLOR_ID_* prefixed definitions from the
+          header include/dt-bindings/leds/common.h. Defaults to

Isn't "include" a Linuxisum?

No, better to have full paths, so automated tools can validate them. If we ever decide to drop it, we can also make a easier search&replace for the pattern starting with include/.

Quoted 15 lines
Quoted 15 lines
Quoted 14 lines
Quoted 11 lines

Documentation/devicetree/bindings/leds/worldsemi,ws2812b.yaml line 57

Quoted 5 lines
Documentation/devicetree/bindings/leds/worldsemi,ws2812b.yaml new file +4 -0
+          <LED_COLOR_ID_GREEN LED_COLOR_ID_RED LED_COLOR_ID_BLUE>
+          if unspecified.
+        $ref: /schemas/types.yaml#/definitions/uint32-array
+        maxItems: 3

In general I am fine with it, although there is still question for adding more multi-color defines in binding headers to replace this property - GRB/RBG/GBR and even more for RGBW.

Pavel, Lee, any thoughts from your side?

This really needs to mention the name this hardware is known as -- I believe it is NeoPixel.

Quoted 7 lines
Quoted 3 lines

We wait here for feedback on colors... The binding is re-implementing color, just because of combinations GRB/RBG/GBR, which could be achieved with new color defines.

Sure, but where does that end?

How many permutations are there likely to be?

For light emitting devices, RGB seems to be used for so long, that I don't expect more permutations (e.g. CMY). On the other hand, someone might create LED strip with whatever colors, so maybe indeed better to allow any variations as in array.

Quoted 1 line

An unlimited array has more of a chance of standing the test of time.

Best regards,
Krzysztof

maintainer
#10 lore raw

Re: [PATCH v4 2/3] dt-bindings: leds: add worldsemi,ws2812b

Quoted 96 lines · Krzysztof Kozlowski
Quoted 90 lines · Lee Jones
Quoted 77 lines · Krzysztof Kozlowski
Quoted 76 lines · Pavel Machek

Hi!

Quoted 74 lines
Quoted 74 lines

Add dt binding schema for WorldSemi WS2812B driven using SPI bus.

---
Changes since v1:
remove linux driver reference from description remove some obvious descriptions
fix unit address regex in multi-led property drop various minItems
add maxItems = 1 to reg
fix node names and property orders in binding example drop -spi from compatible string
add default-brightness

Change since v2:
drop "this patch" from commit message
rename leds to led-controller
drop default-brightness and default-intensity

Change since v3:
reword commit title

 .../bindings/leds/worldsemi,ws2812b.yaml      | 116 ++++++++++++++++++
 1 file changed, 116 insertions(+)
 create mode 100644 Documentation/devicetree/bindings/leds/worldsemi,ws2812b.yaml
Documentation/devicetree/bindings/leds/worldsemi,ws2812b.yaml new file +41 -0
new file mode 100644
index 000000000000..548c05ac3d31
@@ -0,0 +1,116 @@
+# SPDX-License-Identifier: (GPL-2.0-only OR BSD-2-Clause)
+%YAML 1.2
+---
+$id: http://devicetree.org/schemas/leds/worldsemi,ws2812b.yaml#
+$schema: http://devicetree.org/meta-schemas/core.yaml#
+
+title: WS2812B LEDs driven using SPI
+
+maintainers:
+  - Chuanhong Guo <[email protected]>
+
+description: |
+  WorldSemi WS2812B is a individually addressable LED chip that can be chained
+  together and controlled individually using a single wire.
+  This binding describes a chain of WS2812B LEDs connected to the SPI MOSI pin.
+  Typical setups includes connecting the data pin of the LED chain to MOSI as
+  the only device or using CS and MOSI with a tri-state voltage-level shifter
+  for the data pin.
+  The SPI frequency needs to be 2.105MHz~2.85MHz for the timing to be correct
+  and the controller needs to send all the bytes continuously.
+
+allOf:
+  - $ref: /schemas/spi/spi-peripheral-props.yaml#
+
+properties:
+  compatible:
+    const: worldsemi,ws2812b
+
+  reg:
+    maxItems: 1
+
+  spi-max-frequency:
+    minimum: 2105000
+    maximum: 2850000
+
+  "#address-cells":
+    const: 1
+
+  "#size-cells":
+    const: 0
+
Quoted 8 lines
Quoted 8 lines
Quoted 8 lines

Documentation/devicetree/bindings/leds/worldsemi,ws2812b.yaml line 49

Quoted 8 lines
Documentation/devicetree/bindings/leds/worldsemi,ws2812b.yaml new file +8 -0
+patternProperties:
+  "^multi-led@[0-9a-f]+$":
+    type: object
+    $ref: leds-class-multicolor.yaml#
+    unevaluatedProperties: false
+
+    properties:
+      color-index:

Why "index"?

Quoted 1 line

Isn't it just an array of colours rather than an index into something?

Yeah.

Quoted 1 line
Quoted 8 lines
Quoted 6 lines
Quoted 2 lines
Quoted 2 lines
Quoted 2 lines

Documentation/devicetree/bindings/leds/worldsemi,ws2812b.yaml line 51

Quoted 2 lines
Documentation/devicetree/bindings/leds/worldsemi,ws2812b.yaml new file +2 -0
+        description: |
+          A 3-item array specifying color of each components in this LED. It

Why are you forcing this to 3?

Surely there are multi-colour LEDs containing more or less colours?

For this device, because it has only tuples of three.

This doesn't looks like a device specific property to me.

If this is not going to be used by any other device, shouldn't it contain a prefix?

Quoted 8 lines
Quoted 4 lines
Quoted 2 lines
Quoted 2 lines
Quoted 2 lines

Documentation/devicetree/bindings/leds/worldsemi,ws2812b.yaml line 53

Quoted 2 lines
Documentation/devicetree/bindings/leds/worldsemi,ws2812b.yaml new file +2 -0
+          should be one of the LED_COLOR_ID_* prefixed definitions from the
+          header include/dt-bindings/leds/common.h. Defaults to

Isn't "include" a Linuxisum?

No, better to have full paths, so automated tools can validate them. If we ever decide to drop it, we can also make a easier search&replace for the pattern starting with include/.

Very well. It's your train set. :)

Quoted 18 lines
Quoted 18 lines
Quoted 17 lines
Quoted 5 lines
Quoted 5 lines
Quoted 4 lines
Documentation/devicetree/bindings/leds/worldsemi,ws2812b.yaml new file +4 -0
+          <LED_COLOR_ID_GREEN LED_COLOR_ID_RED LED_COLOR_ID_BLUE>
+          if unspecified.
+        $ref: /schemas/types.yaml#/definitions/uint32-array
+        maxItems: 3
Quoted 8 lines
Quoted 5 lines

In general I am fine with it, although there is still question for adding more multi-color defines in binding headers to replace this property - GRB/RBG/GBR and even more for RGBW.

Pavel, Lee, any thoughts from your side?

This really needs to mention the name this hardware is known as -- I believe it is NeoPixel.

We wait here for feedback on colors... The binding is re-implementing color, just because of combinations GRB/RBG/GBR, which could be achieved with new color defines.

Quoted 8 lines
Quoted 3 lines

Sure, but where does that end?

How many permutations are there likely to be?

For light emitting devices, RGB seems to be used for so long, that I don't expect more permutations (e.g. CMY). On the other hand, someone might create LED strip with whatever colors, so maybe indeed better to allow any variations as in array.

Even you suggested variation: "even more for RGBW".

Caveat: as you are well aware, "I'm new here", so my input is no more informed or valuable as yours at this point. I'm just calling it as I see it. If you have strong opinions and they differ wildly from mine, we may have to take intervention from Pavel. As it stands, the property, although slightly restricted IMHO, appears fine.

Quoted 1 line
Quoted 1 line

An unlimited array has more of a chance of standing the test of time.

Signature · 1 line
Lee Jones [李琼斯]
#11 lore raw

Re: [PATCH v4 2/3] dt-bindings: leds: add worldsemi,ws2812b

Hi!

Quoted 94 lines · Lee Jones
Quoted 91 lines · Krzysztof Kozlowski
Quoted 90 lines · Lee Jones
Quoted 77 lines · Krzysztof Kozlowski
Quoted 76 lines · Pavel Machek

Hi!

Quoted 74 lines
Quoted 74 lines

Add dt binding schema for WorldSemi WS2812B driven using SPI bus.

---
Changes since v1:
remove linux driver reference from description remove some obvious descriptions
fix unit address regex in multi-led property drop various minItems
add maxItems = 1 to reg
fix node names and property orders in binding example drop -spi from compatible string
add default-brightness

Change since v2:
drop "this patch" from commit message
rename leds to led-controller
drop default-brightness and default-intensity

Change since v3:
reword commit title

 .../bindings/leds/worldsemi,ws2812b.yaml      | 116 ++++++++++++++++++
 1 file changed, 116 insertions(+)
 create mode 100644 Documentation/devicetree/bindings/leds/worldsemi,ws2812b.yaml
Documentation/devicetree/bindings/leds/worldsemi,ws2812b.yaml new file +41 -0
new file mode 100644
index 000000000000..548c05ac3d31
@@ -0,0 +1,116 @@
+# SPDX-License-Identifier: (GPL-2.0-only OR BSD-2-Clause)
+%YAML 1.2
+---
+$id: http://devicetree.org/schemas/leds/worldsemi,ws2812b.yaml#
+$schema: http://devicetree.org/meta-schemas/core.yaml#
+
+title: WS2812B LEDs driven using SPI
+
+maintainers:
+  - Chuanhong Guo <[email protected]>
+
+description: |
+  WorldSemi WS2812B is a individually addressable LED chip that can be chained
+  together and controlled individually using a single wire.
+  This binding describes a chain of WS2812B LEDs connected to the SPI MOSI pin.
+  Typical setups includes connecting the data pin of the LED chain to MOSI as
+  the only device or using CS and MOSI with a tri-state voltage-level shifter
+  for the data pin.
+  The SPI frequency needs to be 2.105MHz~2.85MHz for the timing to be correct
+  and the controller needs to send all the bytes continuously.
+
+allOf:
+  - $ref: /schemas/spi/spi-peripheral-props.yaml#
+
+properties:
+  compatible:
+    const: worldsemi,ws2812b
+
+  reg:
+    maxItems: 1
+
+  spi-max-frequency:
+    minimum: 2105000
+    maximum: 2850000
+
+  "#address-cells":
+    const: 1
+
+  "#size-cells":
+    const: 0
+
Quoted 8 lines
Quoted 8 lines
Quoted 8 lines

Documentation/devicetree/bindings/leds/worldsemi,ws2812b.yaml line 49

Quoted 8 lines
Documentation/devicetree/bindings/leds/worldsemi,ws2812b.yaml new file +8 -0
+patternProperties:
+  "^multi-led@[0-9a-f]+$":
+    type: object
+    $ref: leds-class-multicolor.yaml#
+    unevaluatedProperties: false
+
+    properties:
+      color-index:

Why "index"?

Quoted 3 lines
Quoted 3 lines
Quoted 1 line

Isn't it just an array of colours rather than an index into something?

Yeah.

The corresponding sysfs interface is called 'multi_index' so I called it this way.

Quoted 10 lines
Quoted 10 lines
Quoted 8 lines
Quoted 4 lines
Quoted 4 lines
Quoted 4 lines

Documentation/devicetree/bindings/leds/worldsemi,ws2812b.yaml line 51

Quoted 4 lines
Documentation/devicetree/bindings/leds/worldsemi,ws2812b.yaml new file +2 -0
+        description: |
+          A 3-item array specifying color of each components in this LED. It

Why are you forcing this to 3?

Surely there are multi-colour LEDs containing more or less colours?

For this device, because it has only tuples of three.

WS2812B has 3 colors per chip. There are chips using a similar protocol with 4 colors but my current driver is hard-coded to support exactly 3 colors.

Quoted 43 lines

This doesn't looks like a device specific property to me.

If this is not going to be used by any other device, shouldn't it contain a prefix?

Quoted 8 lines
Quoted 4 lines
Quoted 2 lines
Quoted 2 lines
Quoted 2 lines

Documentation/devicetree/bindings/leds/worldsemi,ws2812b.yaml line 53

Quoted 2 lines
Documentation/devicetree/bindings/leds/worldsemi,ws2812b.yaml new file +2 -0
+          should be one of the LED_COLOR_ID_* prefixed definitions from the
+          header include/dt-bindings/leds/common.h. Defaults to

Isn't "include" a Linuxisum?

No, better to have full paths, so automated tools can validate them. If we ever decide to drop it, we can also make a easier search&replace for the pattern starting with include/.

Very well. It's your train set. :)

Quoted 18 lines
Quoted 18 lines
Quoted 17 lines
Quoted 5 lines
Quoted 5 lines
Quoted 4 lines
Documentation/devicetree/bindings/leds/worldsemi,ws2812b.yaml new file +4 -0
+          <LED_COLOR_ID_GREEN LED_COLOR_ID_RED LED_COLOR_ID_BLUE>
+          if unspecified.
+        $ref: /schemas/types.yaml#/definitions/uint32-array
+        maxItems: 3
Quoted 8 lines
Quoted 5 lines

In general I am fine with it, although there is still question for adding more multi-color defines in binding headers to replace this property - GRB/RBG/GBR and even more for RGBW.

Pavel, Lee, any thoughts from your side?

This really needs to mention the name this hardware is known as -- I believe it is NeoPixel.

We wait here for feedback on colors... The binding is re-implementing color, just because of combinations GRB/RBG/GBR, which could be achieved with new color defines.

Quoted 8 lines
Quoted 3 lines

Sure, but where does that end?

How many permutations are there likely to be?

For light emitting devices, RGB seems to be used for so long, that I don't expect more permutations (e.g. CMY). On the other hand, someone might create LED strip with whatever colors, so maybe indeed better to allow any variations as in array.

Quoted 9 lines

Even you suggested variation: "even more for RGBW".

Caveat: as you are well aware, "I'm new here", so my input is no more informed or valuable as yours at this point. I'm just calling it as I see it. If you have strong opinions and they differ wildly from mine, we may have to take intervention from Pavel. As it stands, the property, although slightly restricted IMHO, appears fine.

Quoted 1 line
Quoted 1 line

An unlimited array has more of a chance of standing the test of time.

I have another idea that avoids this whole conversation: Abandon color-index completely and determine colors with compatible string and platform data.
My original idea of this property is to support WS2812B and its clones with different colors under the same compatible string. Technically genuine WS2812Bs only come with GRB colors and the clones should have their own chip names (e.g. xiaomi,hm0807a for an RGB clone and <unknown vendor>,sk6812 for an RGBW one.). It's reasonable to have one compatible string per chip for colors, a chip-specific property. Also, adding more compatible devices to a driver is less invasive than adding more definitions to leds/common.h What do you think?

Signature · 2 lines
Regards,
Chuanhong Guo
#12 lore raw

Re: [PATCH v4 2/3] dt-bindings: leds: add worldsemi,ws2812b

Quoted 171 lines · Chuanhong Guo

Hi!

Quoted 94 lines · Lee Jones
Quoted 91 lines · Krzysztof Kozlowski
Quoted 90 lines · Lee Jones
Quoted 77 lines · Krzysztof Kozlowski
Quoted 76 lines · Pavel Machek

Hi!

Quoted 74 lines
Quoted 74 lines

Add dt binding schema for WorldSemi WS2812B driven using SPI bus.

---
Changes since v1:
remove linux driver reference from description remove some obvious descriptions
fix unit address regex in multi-led property drop various minItems
add maxItems = 1 to reg
fix node names and property orders in binding example drop -spi from compatible string
add default-brightness

Change since v2:
drop "this patch" from commit message
rename leds to led-controller
drop default-brightness and default-intensity

Change since v3:
reword commit title

 .../bindings/leds/worldsemi,ws2812b.yaml      | 116 ++++++++++++++++++
 1 file changed, 116 insertions(+)
 create mode 100644 Documentation/devicetree/bindings/leds/worldsemi,ws2812b.yaml
Documentation/devicetree/bindings/leds/worldsemi,ws2812b.yaml new file +41 -0
new file mode 100644
index 000000000000..548c05ac3d31
@@ -0,0 +1,116 @@
+# SPDX-License-Identifier: (GPL-2.0-only OR BSD-2-Clause)
+%YAML 1.2
+---
+$id: http://devicetree.org/schemas/leds/worldsemi,ws2812b.yaml#
+$schema: http://devicetree.org/meta-schemas/core.yaml#
+
+title: WS2812B LEDs driven using SPI
+
+maintainers:
+  - Chuanhong Guo <[email protected]>
+
+description: |
+  WorldSemi WS2812B is a individually addressable LED chip that can be chained
+  together and controlled individually using a single wire.
+  This binding describes a chain of WS2812B LEDs connected to the SPI MOSI pin.
+  Typical setups includes connecting the data pin of the LED chain to MOSI as
+  the only device or using CS and MOSI with a tri-state voltage-level shifter
+  for the data pin.
+  The SPI frequency needs to be 2.105MHz~2.85MHz for the timing to be correct
+  and the controller needs to send all the bytes continuously.
+
+allOf:
+  - $ref: /schemas/spi/spi-peripheral-props.yaml#
+
+properties:
+  compatible:
+    const: worldsemi,ws2812b
+
+  reg:
+    maxItems: 1
+
+  spi-max-frequency:
+    minimum: 2105000
+    maximum: 2850000
+
+  "#address-cells":
+    const: 1
+
+  "#size-cells":
+    const: 0
+
Quoted 8 lines
Quoted 8 lines
Quoted 8 lines

Documentation/devicetree/bindings/leds/worldsemi,ws2812b.yaml line 49

Quoted 8 lines
Documentation/devicetree/bindings/leds/worldsemi,ws2812b.yaml new file +8 -0
+patternProperties:
+  "^multi-led@[0-9a-f]+$":
+    type: object
+    $ref: leds-class-multicolor.yaml#
+    unevaluatedProperties: false
+
+    properties:
+      color-index:

Why "index"?

Quoted 3 lines
Quoted 3 lines
Quoted 1 line

Isn't it just an array of colours rather than an index into something?

Yeah.

The corresponding sysfs interface is called 'multi_index' so I called it this way.

Quoted 10 lines
Quoted 10 lines
Quoted 8 lines
Quoted 4 lines
Quoted 4 lines
Quoted 4 lines

Documentation/devicetree/bindings/leds/worldsemi,ws2812b.yaml line 51

Quoted 4 lines
Documentation/devicetree/bindings/leds/worldsemi,ws2812b.yaml new file +2 -0
+        description: |
+          A 3-item array specifying color of each components in this LED. It

Why are you forcing this to 3?

Surely there are multi-colour LEDs containing more or less colours?

For this device, because it has only tuples of three.

WS2812B has 3 colors per chip. There are chips using a similar protocol with 4 colors but my current driver is hard-coded to support exactly 3 colors.

Quoted 43 lines

This doesn't looks like a device specific property to me.

If this is not going to be used by any other device, shouldn't it contain a prefix?

Quoted 8 lines
Quoted 4 lines
Quoted 2 lines
Quoted 2 lines
Quoted 2 lines

Documentation/devicetree/bindings/leds/worldsemi,ws2812b.yaml line 53

Quoted 2 lines
Documentation/devicetree/bindings/leds/worldsemi,ws2812b.yaml new file +2 -0
+          should be one of the LED_COLOR_ID_* prefixed definitions from the
+          header include/dt-bindings/leds/common.h. Defaults to

Isn't "include" a Linuxisum?

No, better to have full paths, so automated tools can validate them. If we ever decide to drop it, we can also make a easier search&replace for the pattern starting with include/.

Very well. It's your train set. :)

Quoted 18 lines
Quoted 18 lines
Quoted 17 lines
Quoted 5 lines
Quoted 5 lines
Quoted 4 lines
Documentation/devicetree/bindings/leds/worldsemi,ws2812b.yaml new file +4 -0
+          <LED_COLOR_ID_GREEN LED_COLOR_ID_RED LED_COLOR_ID_BLUE>
+          if unspecified.
+        $ref: /schemas/types.yaml#/definitions/uint32-array
+        maxItems: 3
Quoted 8 lines
Quoted 5 lines

In general I am fine with it, although there is still question for adding more multi-color defines in binding headers to replace this property - GRB/RBG/GBR and even more for RGBW.

Pavel, Lee, any thoughts from your side?

This really needs to mention the name this hardware is known as -- I believe it is NeoPixel.

We wait here for feedback on colors... The binding is re-implementing color, just because of combinations GRB/RBG/GBR, which could be achieved with new color defines.

Quoted 8 lines
Quoted 3 lines

Sure, but where does that end?

How many permutations are there likely to be?

For light emitting devices, RGB seems to be used for so long, that I don't expect more permutations (e.g. CMY). On the other hand, someone might create LED strip with whatever colors, so maybe indeed better to allow any variations as in array.

Quoted 9 lines

Even you suggested variation: "even more for RGBW".

Caveat: as you are well aware, "I'm new here", so my input is no more informed or valuable as yours at this point. I'm just calling it as I see it. If you have strong opinions and they differ wildly from mine, we may have to take intervention from Pavel. As it stands, the property, although slightly restricted IMHO, appears fine.

Quoted 1 line
Quoted 1 line

An unlimited array has more of a chance of standing the test of time.

Quoted 12 lines

I have another idea that avoids this whole conversation: Abandon color-index completely and determine colors with compatible string and platform data.
My original idea of this property is to support WS2812B and its clones with different colors under the same compatible string. Technically genuine WS2812Bs only come with GRB colors and the clones should have their own chip names (e.g. xiaomi,hm0807a for an RGB clone and <unknown vendor>,sk6812 for an RGBW one.). It's reasonable to have one compatible string per chip for colors, a chip-specific property. Also, adding more compatible devices to a driver is less invasive than adding more definitions to leds/common.h What do you think?

This sounds good. However it should not lead to compatibles like worldsemi,ws2812b-rgb, worldsemi,ws2812b-bgr etc.

Best regards,
Krzysztof

maintainer
#13 lore raw

Re: [PATCH v4 2/3] dt-bindings: leds: add worldsemi,ws2812b

In reply to Chuanhong Guo
Quoted 97 lines · Chuanhong Guo

Hi!

Quoted 94 lines · Lee Jones
Quoted 91 lines · Krzysztof Kozlowski
Quoted 90 lines · Lee Jones
Quoted 77 lines · Krzysztof Kozlowski
Quoted 76 lines · Pavel Machek

Hi!

Quoted 74 lines
Quoted 74 lines

Add dt binding schema for WorldSemi WS2812B driven using SPI bus.

---
Changes since v1:
remove linux driver reference from description remove some obvious descriptions
fix unit address regex in multi-led property drop various minItems
add maxItems = 1 to reg
fix node names and property orders in binding example drop -spi from compatible string
add default-brightness

Change since v2:
drop "this patch" from commit message
rename leds to led-controller
drop default-brightness and default-intensity

Change since v3:
reword commit title

 .../bindings/leds/worldsemi,ws2812b.yaml      | 116 ++++++++++++++++++
 1 file changed, 116 insertions(+)
 create mode 100644 Documentation/devicetree/bindings/leds/worldsemi,ws2812b.yaml
Documentation/devicetree/bindings/leds/worldsemi,ws2812b.yaml new file +41 -0
new file mode 100644
index 000000000000..548c05ac3d31
@@ -0,0 +1,116 @@
+# SPDX-License-Identifier: (GPL-2.0-only OR BSD-2-Clause)
+%YAML 1.2
+---
+$id: http://devicetree.org/schemas/leds/worldsemi,ws2812b.yaml#
+$schema: http://devicetree.org/meta-schemas/core.yaml#
+
+title: WS2812B LEDs driven using SPI
+
+maintainers:
+  - Chuanhong Guo <[email protected]>
+
+description: |
+  WorldSemi WS2812B is a individually addressable LED chip that can be chained
+  together and controlled individually using a single wire.
+  This binding describes a chain of WS2812B LEDs connected to the SPI MOSI pin.
+  Typical setups includes connecting the data pin of the LED chain to MOSI as
+  the only device or using CS and MOSI with a tri-state voltage-level shifter
+  for the data pin.
+  The SPI frequency needs to be 2.105MHz~2.85MHz for the timing to be correct
+  and the controller needs to send all the bytes continuously.
+
+allOf:
+  - $ref: /schemas/spi/spi-peripheral-props.yaml#
+
+properties:
+  compatible:
+    const: worldsemi,ws2812b
+
+  reg:
+    maxItems: 1
+
+  spi-max-frequency:
+    minimum: 2105000
+    maximum: 2850000
+
+  "#address-cells":
+    const: 1
+
+  "#size-cells":
+    const: 0
+
Quoted 8 lines
Quoted 8 lines
Quoted 8 lines

Documentation/devicetree/bindings/leds/worldsemi,ws2812b.yaml line 49

Quoted 8 lines
Documentation/devicetree/bindings/leds/worldsemi,ws2812b.yaml new file +8 -0
+patternProperties:
+  "^multi-led@[0-9a-f]+$":
+    type: object
+    $ref: leds-class-multicolor.yaml#
+    unevaluatedProperties: false
+
+    properties:
+      color-index:

Why "index"?

Quoted 6 lines
Quoted 3 lines
Quoted 3 lines
Quoted 1 line

Isn't it just an array of colours rather than an index into something?

Yeah.

The corresponding sysfs interface is called 'multi_index' so I called it this way.

DT nomenclature should not be affected by Linuxisums.

Please err towards data-sheet and H/W nomenclature.

Quoted 11 lines
Quoted 8 lines
Quoted 8 lines
Quoted 6 lines
Quoted 2 lines
Quoted 2 lines
Quoted 2 lines

Documentation/devicetree/bindings/leds/worldsemi,ws2812b.yaml line 51

Quoted 2 lines
Documentation/devicetree/bindings/leds/worldsemi,ws2812b.yaml new file +2 -0
+        description: |
+          A 3-item array specifying color of each components in this LED. It

Why are you forcing this to 3?

Surely there are multi-colour LEDs containing more or less colours?

For this device, because it has only tuples of three.

WS2812B has 3 colors per chip. There are chips using a similar protocol with 4 colors but my current driver is hard-coded to support exactly 3 colors.

So the description is for 'this device' rather than any re-use.

And the handling of this property is only contained in your driver?

In which case, my understanding is that you should use a prefix.

Quoted 53 lines
Quoted 43 lines

This doesn't looks like a device specific property to me.

If this is not going to be used by any other device, shouldn't it contain a prefix?

Quoted 8 lines
Quoted 4 lines
Quoted 2 lines
Quoted 2 lines
Quoted 2 lines

Documentation/devicetree/bindings/leds/worldsemi,ws2812b.yaml line 53

Quoted 2 lines
Documentation/devicetree/bindings/leds/worldsemi,ws2812b.yaml new file +2 -0
+          should be one of the LED_COLOR_ID_* prefixed definitions from the
+          header include/dt-bindings/leds/common.h. Defaults to

Isn't "include" a Linuxisum?

No, better to have full paths, so automated tools can validate them. If we ever decide to drop it, we can also make a easier search&replace for the pattern starting with include/.

Very well. It's your train set. :)

Quoted 18 lines
Quoted 18 lines
Quoted 17 lines
Quoted 5 lines
Quoted 5 lines
Quoted 4 lines
Documentation/devicetree/bindings/leds/worldsemi,ws2812b.yaml new file +4 -0
+          <LED_COLOR_ID_GREEN LED_COLOR_ID_RED LED_COLOR_ID_BLUE>
+          if unspecified.
+        $ref: /schemas/types.yaml#/definitions/uint32-array
+        maxItems: 3
Quoted 8 lines
Quoted 5 lines

In general I am fine with it, although there is still question for adding more multi-color defines in binding headers to replace this property - GRB/RBG/GBR and even more for RGBW.

Pavel, Lee, any thoughts from your side?

This really needs to mention the name this hardware is known as -- I believe it is NeoPixel.

We wait here for feedback on colors... The binding is re-implementing color, just because of combinations GRB/RBG/GBR, which could be achieved with new color defines.

Quoted 8 lines
Quoted 3 lines

Sure, but where does that end?

How many permutations are there likely to be?

For light emitting devices, RGB seems to be used for so long, that I don't expect more permutations (e.g. CMY). On the other hand, someone might create LED strip with whatever colors, so maybe indeed better to allow any variations as in array.

Quoted 9 lines

Even you suggested variation: "even more for RGBW".

Caveat: as you are well aware, "I'm new here", so my input is no more informed or valuable as yours at this point. I'm just calling it as I see it. If you have strong opinions and they differ wildly from mine, we may have to take intervention from Pavel. As it stands, the property, although slightly restricted IMHO, appears fine.

Quoted 1 line
Quoted 1 line

An unlimited array has more of a chance of standing the test of time.

Quoted 12 lines

I have another idea that avoids this whole conversation: Abandon color-index completely and determine colors with compatible string and platform data.
My original idea of this property is to support WS2812B and its clones with different colors under the same compatible string. Technically genuine WS2812Bs only come with GRB colors and the clones should have their own chip names (e.g. xiaomi,hm0807a for an RGB clone and <unknown vendor>,sk6812 for an RGBW one.). It's reasonable to have one compatible string per chip for colors, a chip-specific property. Also, adding more compatible devices to a driver is less invasive than adding more definitions to leds/common.h What do you think?

Yes it's reasonable to have a compatible string per device and yes, you can do matching (in C) based on compatible string, so this sounds reasonable to me. It also sounds reasonable to describe the H/W (inc. colours along with any ordering, if that's important) inside the DT.

Your call I guess.

Signature · 1 line
Lee Jones [李琼斯]
#14 lore raw

Re: [PATCH v4 2/3] dt-bindings: leds: add worldsemi,ws2812b

Hi!

Quoted 6 lines · Lee Jones

[...]
So the description is for 'this device' rather than any re-use.

And the handling of this property is only contained in your driver?

In which case, my understanding is that you should use a prefix.

(Just out of curiosity. I don't want this property now.)

My understanding is that a vendor prefix means this property is for devices by this vendor. However color-index is for supporting clones, which are chips not made by this vendor. Does a vendor prefix really apply in this case?

--
Regards,
Chuanhong Guo

maintainer
#15 lore raw

Re: [PATCH v4 2/3] dt-bindings: leds: add worldsemi,ws2812b

Quoted 8 lines · Chuanhong Guo

Hi!

Quoted 5 lines · Lee Jones

[...]
So the description is for 'this device' rather than any re-use.

And the handling of this property is only contained in your driver?

Quoted 8 lines
Quoted 1 line

In which case, my understanding is that you should use a prefix.

(Just out of curiosity. I don't want this property now.)

My understanding is that a vendor prefix means this property is for devices by this vendor. However color-index is for supporting clones, which are chips not made by this vendor. Does a vendor prefix really apply in this case?

No idea. Sounds like a grey area.

Signature · 1 line
Lee Jones [李琼斯]
#16 lore raw

PATCH v4 3/3 leds: add driver for SPI driven WorldSemi WS2812B RGB LEDs

In reply to Chuanhong Guo

This patch adds support for driving a chain of WS2812B LED chips using SPI bus.

WorldSemi WS2812B is a individually addressable LED chip that can be chained together and controlled individually using a single wire. The chip recognize a long pulse as a bit of 1 and a short pulse as a bit of 0. Host sends a continuous stream of 24-bits color values, each LED chip takes the first 3 byte it receives as its color value and passes the leftover bytes to the next LED on the chain.

This driver simulates this protocol using SPI bus by sending a long pulse as 3'b110 and a short pulse as 3'b100. The SPI frequency needs to be 2.105MHz~2.85MHz for the timing to be correct and the controller needs to transfer all the bytes continuously.

---
Changes since v1:
rename the driver to drop -spi suffix
add support for default-brightness
use fwnode apis for properties

Changes since v2:
drop default-brightness and default-intensity

Changes since v3:
1. add more comments
2. rename reg to cascade
3. redo some line breaking
4. move duplicated pointer calculation into ws2812b_set_byte
5. reword error message
6. get ws2812b_priv from led cdev->dev->parent

 drivers/leds/rgb/Kconfig        |  11 ++
 drivers/leds/rgb/Makefile       |   1 +
 drivers/leds/rgb/leds-ws2812b.c | 231 ++++++++++++++++++++++++++++++++
 3 files changed, 243 insertions(+)
 create mode 100644 drivers/leds/rgb/leds-ws2812b.c
drivers/leds/rgb/Kconfig +11 -0
index 204cf470beae..5c2081852f01 100644
@@ -26,4 +26,15 @@ config LEDS_QCOM_LPG
 
 	  If compiled as a module, the module will be named leds-qcom-lpg.
 
+config LEDS_WS2812B
+	tristate "SPI driven WS2812B RGB LED support"
+	depends on OF
+	depends on SPI
+	help
+	  This option enables support for driving daisy-chained WS2812B RGB
+	  LED chips using SPI bus. This driver simulates the single-wire
+	  protocol by sending bits over the SPI MOSI pin. For this to work,
+	  the SPI frequency should be 2.105MHz~2.85MHz and the controller
+	  needs to transfer all the bytes continuously.
+
 endif # LEDS_CLASS_MULTICOLOR
drivers/leds/rgb/Makefile +1 -0
index 0675bc0f6e18..a6f855eaeb14 100644
@@ -2,3 +2,4 @@
 
 obj-$(CONFIG_LEDS_PWM_MULTICOLOR)	+= leds-pwm-multicolor.o
 obj-$(CONFIG_LEDS_QCOM_LPG)		+= leds-qcom-lpg.o
+obj-$(CONFIG_LEDS_WS2812B)		+= leds-ws2812b.o
drivers/leds/rgb/leds-ws2812b.c new file +231 -0

231 lines not on this page · raw message

Signature · 1 line
2.38.1