[PATCH 0/4] Add I2C support on TH1520

[Thomas Bonnefille]

On 4/26/24 4:42 PM, Conor Dooley wrote:
> On Fri, Apr 26, 2024 at 04:12:00PM +0200, Thomas Bonnefille wrote:
>>
>>
>> On 4/25/24 6:35 PM, Conor Dooley wrote:
>>> On Thu, Apr 25, 2024 at 09:51:26AM -0500, Rob Herring wrote:
>>>>
>>>> On Thu, 25 Apr 2024 10:21:31 +0200, Thomas Bonnefille wrote:
>>>>> This adds I2C support in the device tree of the T-Head TH1520 RISCV-SoC
>>>>> and a default configuration for the BeagleV-Ahead. It appears that the
>>>>> TH1520 I2C is already supported in the upstream kernel through the
>>>>> Synopsis Designware I2C adapter driver.
>>>>> As there is no clock driver for this board as of today, this patch
>>>>> series uses a fixed-clock named i2c_ic_clk.
>>>>> There is also no pinctrl driver yet so pinmux must be handled manually
>>>>> for now.
>>>>> It also fixes the order of the nodes in the device tree to comply with
>>>>> device-tree coding-style.
>>>>>
>>>>> Thomas Bonnefille (4):
>>>>>     dt-bindings: i2c: dw: Document compatible thead,th1520-i2c
>>>>>     riscv: boot: dts: thead: Fix node ordering in TH1520 device tree
>>>>>     riscv: dts: thead: Add TH1520 I2C nodes
>>>>>     riscv: dts: thead: Enable I2C on the BeagleV-Ahead
>>>>>
>>>>>    .../bindings/i2c/snps,designware-i2c.yaml     |  12 ++
>>>>>    .../boot/dts/thead/th1520-beaglev-ahead.dts   |  22 ++++
>>>>>    arch/riscv/boot/dts/thead/th1520.dtsi         | 120 ++++++++++++++----
>>>>>    3 files changed, 127 insertions(+), 27 deletions(-)
>>>>>
>>>>> --
>>>>> 2.44.0
>>>>>
>>>>>
>>>>>
>>>>
>>>>
>>>> My bot found new DTB warnings on the .dts files added or changed in this
>>>> series.
>>>>
>>>> Some warnings may be from an existing SoC .dtsi. Or perhaps the warnings
>>>> are fixed by another series. Ultimately, it is up to the platform
>>>> maintainer whether these warnings are acceptable or not. No need to reply
>>>> unless the platform maintainer has comments.
>>>>
>>>> If you already ran DT checks and didn't see these error(s), then
>>>> make sure dt-schema is up to date:
>>>>
>>>>     pip3 install dtschema --upgrade
>>>>
>>>>
>>>> New warnings running 'make CHECK_DTBS=y thead/th1520-beaglev-ahead.dtb' for 20240425082138.374445-1-thomas.bonnefille at bootlin.com:
>>>>
>>>> arch/riscv/boot/dts/thead/th1520-beaglev-ahead.dtb: i2c-clock: clock-frequency:0:0: 50000000 is greater than the maximum of 5000000
>>>
>>> The bot is not freaking out here, 50 MHz is indeed more than 5 MHz :)
>>
>> 5MHz is the maximum clock-frequency, ie. the I2C bus frequency.
>> This is actually set to 100kHz for I2C0 in the DT:
>>
>> &i2c0 {
>>      status = "okay";
>>      clock-frequency = <100000>; <----
>> ...
>> };
>>
>> 50MHz is the "fixed-clock" frequency, that is the clock feeding the I2C
>> IP block:
>>
>> i2c_ic_clk: i2c-clock {
>>      compatible = "fixed-clock";
>>      clock-frequency = <50000000>; <-----
>>      #clock-cells = <0>;
>> };
>>
>> My guess is that the bot confused the clock-frequency parameter for the bus
>> clock (SCL) with the i2c-ic-clock value for the controller itself during the
>> checks.
>>
>> Do you agree with this or am I misunderstanding the error ?
>>
>> If I lower the fixed-clock frequency to eg. 100kHz, the error is gone. But I
>> guess the 5MHz limit should probably not apply to the input clock?
> 
> Heh, I know why that's happening - it's your node name.
> The pattern for i2c controllers is "^i2c(@.*)?":
> https://github.com/devicetree-org/dt-schema/blob/main/dtschema/schemas/i2c/i2c-controller.yaml
> 
> Rob co-incidentally (or maybe not) put out a patch for fixed-frequency
> clock names, suggesting using clock as a prefix:
> https://lore.kernel.org/all/20240425183810.3079069-1-robh@kernel.org/
> If we switched to that format, I believe your problem goes away.
> 
> Cheers,
> Conor.
> 

Indeed it was the node name, when I switched the name of the fixed clock 
to clock-i2c-ic, the error was gone, thank you. :)
But i2c_ic_clk shouldn't match "^i2c(@.*)?", one of my teammates 
suggested that the error may instead come from this line 
https://elixir.bootlin.com/linux/latest/source/scripts/dtc/checks.c#L1023.

But as Drew is working on new iterations of the TH1520's clock driver I 
will just delete this fixed clock and publish a second version depending 
on his patch.

Thank you,
Thomas