Adding Benjamin who mainly implemented this.

On Tue, May 05, 2020 at 07:51:08AM +0200, Alain Volmat wrote:
> SMBus Host-Notify protocol, from the adapter point of view
> consist of receiving a message from a client, including the
> client address and some other data.
> 
> It can be simply handled by creating a new slave device
> and registering a callback performing the parsing of the
> message received from the client.
> 
> This commit introduces two new core functions
>   * i2c_new_smbus_host_notify_device
>   * i2c_free_smbus_host_notify_device
> that take care of registration of the new slave device and
> callback and will call i2c_handle_smbus_host_notify once a
> Host-Notify event is received.

Yay, cool idea to use the slave interface. I like it a lot!

> +static int i2c_smbus_host_notify_cb(struct i2c_client *client,
> +				    enum i2c_slave_event event, u8 *val)
> +{
> +	struct i2c_smbus_host_notify_status *status = client->dev.platform_data;
> +	int ret;
> +
> +	switch (event) {
> +	case I2C_SLAVE_WRITE_REQUESTED:
> +		status->notify_start = true;
> +		break;
> +	case I2C_SLAVE_WRITE_RECEIVED:
> +		/* We only retrieve the first byte received (addr)
> +		 * since there is currently no way to retrieve the data
> +		 * parameter from the client.

Maybe s/no way/no support/ ? I still wonder if we couldn't add it
somehow. Once we find a device which needs this, of course.

> +		 */
> +		if (!status->notify_start)
> +			break;
> +		status->addr = *val;
> +		status->notify_start = false;
> +		break;
> +	case I2C_SLAVE_STOP:

What about setting 'notify_start' to false here as well? In the case of
an incomplete write?

> +		ret = i2c_handle_smbus_host_notify(client->adapter,
> +						   status->addr);
> +		if (ret < 0) {
> +			dev_warn(&client->adapter->dev, "failed to handle host_notify (%d)\n",
> +				ret);

I think we should rather add such error strings to the core if we think
they are needed. I am not convinced they are, though.

> +			return ret;
> +		}
> +		break;
> +	default:
> +		/* Only handle necessary events */
> +		break;
> +	}
> +
> +	return 0;
> +}
> +

Rest of the code looks good. Maybe we should compile all this only when
I2C_SLAVE is enabled?