Re: [PATCH v1 1/3] dt-bindings: i2c: Add Qualcomm I2C slave controller

[Mukesh Savaliya <mukesh.savaliya@oss.qualcomm.com>]

On 6/29/2026 11:56 AM, Wolfram Sang wrote:
> Hi,
> 
> On Sun, Jun 28, 2026 at 08:09:04PM +0530, Viken Dadhaniya wrote:
>> Add a binding for the Qualcomm I2C slave controller found on QDU1000. The
>> block operates only as an I2C slave and supports FIFO/PIO transfers, so it
>> needs a binding separate from Qualcomm I2C master controllers.
> 
> The official naming in I2C is now 'target' instead of 'slave'. Do you
> think you can rename that? I know the Linux I2C core has still the old
> wording in a lot of places and needs to be converted. However, if
> possible, it would be nice if not more of the old terms would be added
> which need to be converted at some point anyhow.
> 
Agree, let's change to target naming.
>> Document the MMIO region, interrupt, XO and AHB clocks, interconnect path,
>> and the 7-bit slave address needed to describe the controller in device
>> tree.
>>
>> Use the 'qcom,slave-addr' property for the slave address because 'reg'
>> describes the controller MMIO range, and this slave-only controller has no
>> child node where an I2C address can be encoded.
> 
> This should not be needed because the backend defines which address is
> going to be used? Can you share how you tested this?
> 
> Happy hacking,
> 
>     Wolfram
>