On 08/11/2022 21:38, Andrew Lunn wrote:
>> +static int h4_probe(struct serdev_device *serdev)
>> +{
>> + struct h4_device *h4dev;
>> + struct hci_uart *hu;
>> + int ret;
>> + u32 speed;
>> +
>> + h4dev = devm_kzalloc(&serdev->dev, sizeof(*h4dev), GFP_KERNEL);
>> + if (!h4dev)
>> + return -ENOMEM;
>> +
>> + hu = &h4dev->hu;
>> +
>> + hu->serdev = serdev;
>> + ret = device_property_read_u32(&serdev->dev, "max-speed", &speed);
>> + if (!ret) {
>> + /* h4 does not have any baudrate handling:
>> + * user oper speed from the start
>> + */
>> + hu->init_speed = speed;
>> + hu->oper_speed = speed;
>> + }
>> +
>> + serdev_device_set_drvdata(serdev, h4dev);
>> + dev_info(&serdev->dev, "h4 device registered.\n");
>
> It is considered bad practice to spam the logs like this. dev_dbg().
>
>> +
>> + return hci_uart_register_device(hu, &h4p);
>> +}
>> +
>> +static void h4_remove(struct serdev_device *serdev)
>> +{
>> + struct h4_device *h4dev = serdev_device_get_drvdata(serdev);
>> +
>> + dev_info(&serdev->dev, "h4 device unregistered.\n");
>
> dev_dbg().
I would say none of them (the same in probe). Any prints in probe/remove
paths are considered redundant, as core already gives that information.
Best regards,
Krzysztof
