Skip to contentSkip to conversation

dt-bindings: i2c: Remove obsolete i2c.txt

PATCH In review Rob Herring 3 participants 7 messages Feb 22, 2024 – Feb 27, 2024

Subsystems (7)
Device tree bindings Sound ASoC GPIO +3 more subsystems
lore
Share & tools
lore

Rob Herring

author maintainer
#0 lore raw

PATCH 1/1 dt-bindings: i2c: Remove obsolete i2c.txt

Everything in i2c.txt is covered by schemas/i2c/i2c-controller.yaml in dtschema project, so remove i2c.txt and update links to it in the tree.

---
Wolfram, you can take it or I can.

 .../bindings/gpio/gateworks,pld-gpio.txt      |   3 +-
 Documentation/devicetree/bindings/i2c/i2c.txt | 151 ------------------
 .../i2c/nvidia,tegra186-bpmp-i2c.yaml         |   3 +-
 .../devicetree/bindings/i3c/i3c.yaml          |   2 +-
 .../devicetree/bindings/sound/cs4341.txt      |   2 +-
 MAINTAINERS                                   |   1 -
 6 files changed, 4 insertions(+), 158 deletions(-)
 delete mode 100644 Documentation/devicetree/bindings/i2c/i2c.txt
Documentation/devicetree/bindings/gpio/gateworks,pld-gpio.txt +1 -2
index 6e81f8b755c5..d543fd1b8b23 100644
@@ -1,7 +1,6 @@
 Gateworks PLD GPIO controller bindings
 
-The GPIO controller should be a child node on an I2C bus,
-see: i2c/i2c.txt for details.
+The GPIO controller should be a child node on an I2C bus.
 
 Required properties:
 - compatible: Should be "gateworks,pld-gpio"
Documentation/devicetree/bindings/i2c/i2c.txt deleted +0 -151

151 lines not on this page · raw message

Documentation/devicetree/bindings/i2c/nvidia,tegra186-bpmp-i2c.yaml +1 -2
index b8319dcf3d8a..8676335e9e94 100644
@@ -21,8 +21,7 @@ description: |
   See ../firmware/nvidia,tegra186-bpmp.yaml for details of the BPMP
   binding.
 
-  This node represents an I2C controller. See ../i2c/i2c.txt for details
-  of the core I2C binding.
+  This node represents an I2C controller.
 
 properties:
   compatible:
Documentation/devicetree/bindings/i3c/i3c.yaml +1 -1
index c816e295d565..87cadbcdc61c 100644
@@ -71,7 +71,7 @@ patternProperties:
     description: |
       I2C child, should be named: <device-type>@<i2c-address>
 
-      All properties described in Documentation/devicetree/bindings/i2c/i2c.txt
+      All properties described in dtschema schemas/i2c/i2c-controller.yaml
       are valid here, except the reg property whose content is changed.
 
     properties:
Documentation/devicetree/bindings/sound/cs4341.txt +1 -1
index 12b4aa8ef0db..c1d5c8ad1a36 100644
@@ -9,7 +9,7 @@ Required properties:
           number for SPI.
 
 For required properties on I2C-bus, please consult
-Documentation/devicetree/bindings/i2c/i2c.txt
+dtschema schemas/i2c/i2c-controller.yaml
 For required properties on SPI-bus, please consult
 Documentation/devicetree/bindings/spi/spi-bus.txt
 
MAINTAINERS +0 -1
index 45c6c13b4edf..50a906eb8dfd 100644
@@ -10074,7 +10074,6 @@ S:	Maintained
 W:	https://i2c.wiki.kernel.org/
 Q:	https://patchwork.ozlabs.org/project/linux-i2c/list/
 T:	git git://git.kernel.org/pub/scm/linux/kernel/git/wsa/linux.git
-F:	Documentation/devicetree/bindings/i2c/i2c.txt
 F:	Documentation/i2c/
 F:	drivers/i2c/*
 F:	include/dt-bindings/i2c/i2c.h
Signature · 1 line
2.43.0
maintainer Acked-by on v1 1/1
#1 lore raw
Quoted 2 lines · Rob Herring

Everything in i2c.txt is covered by schemas/i2c/i2c-controller.yaml in dtschema project, so remove i2c.txt and update links to it in the tree.

#2 lore raw
In reply to Rob Herring
Quoted 4 lines · Rob Herring

Everything in i2c.txt is covered by schemas/i2c/i2c-controller.yaml in dtschema project, so remove i2c.txt and update links to it in the tree.

Differences to i2c.txt:

* In the schema, "clock-frequency" has a minimum of 1kHz and a maximum of 3MHz. Why? The specs do not say anything about a minimum freq and fastest speed mentioned in the docs is 5Mhz (Ultra fast mode).

* new binding "i2c-scl-clk-low-timeout-us" has a description which I do not understand. What is a waiting state?

* new binding "no-detect" is broken. At the least, it should be named something like "bus-fully-described" and then the OS can decide to leave out auto-detection mechanisms. If you are interested in the latter, you can simply disable class based instantiation on the host controller. No need to describe this in DT.

Quoted 1 line

Wolfram, you can take it or I can.

Once we are done, I guess it is better if you take it.

Thanks for the work,

Wolfram

author maintainer
#3 lore raw
Quoted 8 lines · Wolfram Sang
Quoted 4 lines · Rob Herring

Everything in i2c.txt is covered by schemas/i2c/i2c-controller.yaml in dtschema project, so remove i2c.txt and update links to it in the tree.

Differences to i2c.txt:

Thanks

Quoted 3 lines

* In the schema, "clock-frequency" has a minimum of 1kHz and a maximum of 3MHz. Why? The specs do not say anything about a minimum freq and fastest speed mentioned in the docs is 5Mhz (Ultra fast mode).

IIRC, the high speed mode originally topped out at 3MHz. I guess that's been revised.

We can drop the minimum.

Quoted 2 lines

* new binding "i2c-scl-clk-low-timeout-us" has a description which I do not understand. What is a waiting state?

Shrug. May have to look at the MPC h/w that uses the property.

Quoted 6 lines

* new binding "no-detect" is broken. At the least, it should be named something like "bus-fully-described" and then the OS can decide to leave out auto-detection mechanisms. If you are interested in the latter, you can simply disable class based instantiation on the host controller. No need to describe this in DT.

I've reverted the property now.

Rob

#4 lore raw

Hi Rob,

Quoted 6 lines
Quoted 3 lines

* In the schema, "clock-frequency" has a minimum of 1kHz and a maximum of 3MHz. Why? The specs do not say anything about a minimum freq and fastest speed mentioned in the docs is 5Mhz (Ultra fast mode).

IIRC, the high speed mode originally topped out at 3MHz. I guess that's been revised.

Hs-mode has a max of 3.4MHz...

Quoted 2 lines

We can drop the minimum.

... but I see you changed min/max now to 1/5000000. That's what I would have suggested as well.

Quoted 4 lines
Quoted 2 lines

* new binding "i2c-scl-clk-low-timeout-us" has a description which I do not understand. What is a waiting state?

Shrug. May have to look at the MPC h/w that uses the property.

I will also have another look. My gut feeling is that the binding is okay, only the description might need an update.

Quoted 8 lines
Quoted 6 lines

* new binding "no-detect" is broken. At the least, it should be named something like "bus-fully-described" and then the OS can decide to leave out auto-detection mechanisms. If you are interested in the latter, you can simply disable class based instantiation on the host controller. No need to describe this in DT.

I've reverted the property now.

Cool, thanks!

Kind regards,

Wolfram

author maintainer
#5 lore raw
Quoted 34 lines · Wolfram Sang

Hi Rob,

Quoted 6 lines
Quoted 3 lines

* In the schema, "clock-frequency" has a minimum of 1kHz and a maximum of 3MHz. Why? The specs do not say anything about a minimum freq and fastest speed mentioned in the docs is 5Mhz (Ultra fast mode).

IIRC, the high speed mode originally topped out at 3MHz. I guess that's been revised.

Hs-mode has a max of 3.4MHz...

Quoted 2 lines

We can drop the minimum.

... but I see you changed min/max now to 1/5000000. That's what I would have suggested as well.

Quoted 4 lines
Quoted 2 lines

* new binding "i2c-scl-clk-low-timeout-us" has a description which I do not understand. What is a waiting state?

Shrug. May have to look at the MPC h/w that uses the property.

I will also have another look. My gut feeling is that the binding is okay, only the description might need an update.

Quoted 7 lines
Quoted 6 lines

* new binding "no-detect" is broken. At the least, it should be named something like "bus-fully-described" and then the OS can decide to leave out auto-detection mechanisms. If you are interested in the latter, you can simply disable class based instantiation on the host controller. No need to describe this in DT.

Quoted 3 lines
Quoted 1 line

I've reverted the property now.

Cool, thanks!

I don't think there's anything else to discuss on this patch, can I get your ack?

Rob