Re: [PATCH] i2c: Devices which have some i2c addr can work in same i2c bus

[Andi Shyti <andi.shyti@kernel.org>]

Hi David,

On Thu, Jun 01, 2023 at 11:34:23AM +0800, David Wu wrote:
> From: Zhang aihui <zah@rock-chips.com>
> 
> If i2c slave devices don't work at the same time, which have
> the same i2c addr, it would register two devices, can make them
> working.

can you please rephrase this?

I understand you want to register multiple devices, how is this
going to work in hardware?

> Change-Id: I1bfb7783924b08bdc6e12bf47c2de01bdac7c2e2

please drop the Change-Id

> Signed-off-by: Zhang aihui <zah@rock-chips.com>
> Signed-off-by: David Wu <david.wu@rock-chips.com>
> ---
>  drivers/i2c/i2c-core-base.c | 51 +++++++++++++++++++++++++++++--------
>  1 file changed, 41 insertions(+), 10 deletions(-)
> 
> diff --git a/drivers/i2c/i2c-core-base.c b/drivers/i2c/i2c-core-base.c
> index ae3af738b03f..53a8141e6238 100644
> --- a/drivers/i2c/i2c-core-base.c
> +++ b/drivers/i2c/i2c-core-base.c
> @@ -62,6 +62,7 @@
>  static DEFINE_MUTEX(core_lock);
>  static DEFINE_IDR(i2c_adapter_idr);
>  
> +static int i2c_check_addr_ex(struct i2c_adapter *adapter, int addr);
>  static int i2c_detect(struct i2c_adapter *adapter, struct i2c_driver *driver);
>  
>  static DEFINE_STATIC_KEY_FALSE(i2c_trace_msg_key);
> @@ -849,7 +850,8 @@ static void i2c_adapter_unlock_bus(struct i2c_adapter *adapter,
>  
>  static void i2c_dev_set_name(struct i2c_adapter *adap,
>  			     struct i2c_client *client,
> -			     struct i2c_board_info const *info)
> +			     struct i2c_board_info const *info,
> +			     int status)

what exactly is status, is it a counter? If so, please call it
count or similar.

>  {
>  	struct acpi_device *adev = ACPI_COMPANION(&client->dev);
>  
> @@ -863,8 +865,12 @@ static void i2c_dev_set_name(struct i2c_adapter *adap,
>  		return;
>  	}
>  
> -	dev_set_name(&client->dev, "%d-%04x", i2c_adapter_id(adap),
> -		     i2c_encode_flags_to_addr(client));
> +	if (status == 0)
> +		dev_set_name(&client->dev, "%d-%04x", i2c_adapter_id(adap),
> +			i2c_encode_flags_to_addr(client));
> +	else
> +		dev_set_name(&client->dev, "%d-%04x-%01x", i2c_adapter_id(adap),
> +			i2c_encode_flags_to_addr(client), status);
>  }
>  
>  int i2c_dev_irq_from_resources(const struct resource *resources,
> @@ -940,9 +946,11 @@ i2c_new_client_device(struct i2c_adapter *adap, struct i2c_board_info const *inf
>  	}
>  
>  	/* Check for address business */
> -	status = i2c_check_addr_busy(adap, i2c_encode_flags_to_addr(client));
> +	status = i2c_check_addr_ex(adap, i2c_encode_flags_to_addr(client));

"status" as such was indicating that the device is busy, i.e. the
device exists. If you want to use it as a counter, then make
another variable, u8, possibly.

>  	if (status)
> -		goto out_err;
> +		dev_err(&adap->dev,
> +			"%d i2c clients have been registered at 0x%02x",

I think rather than dev_err() it should be a dev_warn() (or
dev_info() as the message doesn't sound very threatening).

dev_err() should be normally followed by a failure. Perhaps to
make it sound more as a warning the message should be:

	"client %d is already registere in 0x%02x\n"

Andi

> +			status, client->addr);