!usbd_register_set_config_callback - no handler copy

[alexrayne]

just a small fix protects register one callback multiple times, that will loose callbacl slots and may prevent executing configuration multiple times at activation

[karlp]

So to clarify, this is intended to prevent registering the same callback repeatedly?

[alexrayne]

yep, it is intended for implementing composit devices with multiple functions - if i create multile MSC interfaces, they can try to register ones callback per every interface. Without this patch, i must perform registration of functional class, outside of class functionality-implementer.
Maybe this aproach more careful, but this patch gives more fredom for library users.

[ChuckM]

I would write this differently

if ((usbd_dev->user_callback_set_config[i] != NULL) &&
   (usbd_dev->user_callback_set_config[i] == callback)) {
    return 0;
}

The multiple continues in there are a bit weird.

[karlp]

it's so that it continues properly with the right return codes though.

[ChuckM]

You are right @karlp here is (for me) the more readable and understandable code:

int usbd_register_set_config_callback(usbd_device *usbd_dev,
				       usbd_set_config_callback callback)
{
	int i;
	for (i = 0; i < MAX_USER_SET_CONFIG_CALLBACK; i++) {
		if ((usbd_dev->user_callback_set_config[i] == NULL) ||
                    (usbd_dev->user_callback_set_config[i] == callback)) {
			usbd_dev->user_callback_set_config[i] = callback;
                        return 0;
               }
        }
       return -1;
}

[karlp]

hehe :) I'd worked up this version myself locally:

	for (i = 0; i < MAX_USER_SET_CONFIG_CALLBACK; i++) {
		if (usbd_dev->user_callback_set_config[i]) {
			if (usbd_dev->user_callback_set_config[i] == callback) {
				return 0;
			}
			continue;
		}

		usbd_dev->user_callback_set_config[i] = callback;
		return 0;
	}

[karlp]

an implementation of this has now merged, see f4bbe7c