On Thu, Jun 25, 2020 at 09:39:26AM +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.
> 
> Signed-off-by: Alain Volmat <alain.volmat@st.com>
> ---
> v2: remove useless dev_err message in case of hnotify handling error
>     prevent handling hnotify in case of a incomplete write

Okay, now I got it to work, I also noted a few more issues.

First, I'd suggest s/i2c_smbus_host_notify/i2c_slave_host_notify/g for
all occurences in this patch. This makes a stronger distinction between
the generic HostNotify support and the slave specific one.

Also, I wonder if this shouldn't go to i2c-smbus.c instead but I haven't
checked if we end up in dependency hell then. Second best thought: at
least move to i2c-core-slave.c, then we could save the #ifdeffery in the
c-file?

> 
>  drivers/i2c/i2c-core-smbus.c | 110 +++++++++++++++++++++++++++++++++++++++++++
>  include/linux/i2c-smbus.h    |   2 +
>  2 files changed, 112 insertions(+)
> 
> diff --git a/drivers/i2c/i2c-core-smbus.c b/drivers/i2c/i2c-core-smbus.c
> index 56bb840142e3..3a37664fb5f6 100644
> --- a/drivers/i2c/i2c-core-smbus.c
> +++ b/drivers/i2c/i2c-core-smbus.c
> @@ -708,3 +708,113 @@ int of_i2c_setup_smbus_alert(struct i2c_adapter *adapter)
>  }
>  EXPORT_SYMBOL_GPL(of_i2c_setup_smbus_alert);
>  #endif
> +
> +#if IS_ENABLED(CONFIG_I2C_SLAVE)
> +struct i2c_smbus_host_notify_status {
> +	bool notify_start;
> +	u8 addr;
> +};
> +
> +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 support to retrieve the data
> +		 * parameter from the client.
> +		 */
> +		if (!status->notify_start)
> +			break;
> +		status->addr = *val;
> +		status->notify_start = false;

So, we are safe if the message is too short. Otherwise, we capture the
first byte (== address) only, right. Further bytes until STOP are
discarded. So, we don't check if the message is too long and contains
more than the status word. Maybe we should add that?

> +		break;
> +	case I2C_SLAVE_STOP:
> +		/* In case of incomplete write, don't handle host-notify */
> +		if (status->notify_start) {
> +			status->notify_start = false;
> +			break;
> +		}
> +
> +		ret = i2c_handle_smbus_host_notify(client->adapter,
> +						   status->addr);
> +		if (ret < 0)
> +			return ret;
> +		break;
> +	default:

The missing cases are mandatory. From my testunit driver:

        case I2C_SLAVE_READ_REQUESTED:
        case I2C_SLAVE_READ_PROCESSED:
                *val = 0xff;
                break;

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

...

> --- a/include/linux/i2c-smbus.h
> +++ b/include/linux/i2c-smbus.h
> @@ -38,6 +38,8 @@ static inline int of_i2c_setup_smbus_alert(struct i2c_adapter *adap)
>  	return 0;
>  }
>  #endif
> +struct i2c_client *i2c_new_smbus_host_notify_device(struct i2c_adapter *adapter);
> +void i2c_free_smbus_host_notify_device(struct i2c_client *client);

Those need to be guarded with I2C_SLAVE as well. And an #else branch
with empty/successful placeholders.

>  
>  #if IS_ENABLED(CONFIG_I2C_SMBUS) && IS_ENABLED(CONFIG_DMI)
>  void i2c_register_spd(struct i2c_adapter *adap);
> -- 
> 2.7.4
>