Hi Johan, >>> If either one of the HCI_SETUP or HCI_USER_CHANNEL flags is set the >>> device is not considered valid for mgmt. By having these checks inside >>> the mgmt_valid_hdev function the a couple of places using it can be >>> simplified. >> >> I looked at doing this and decided not to. Reason was that the device >> gets removed from mgmt anyway. > > I'm not sure what significance you think "device gets removed from mgmt" > has. All that is is a mgmt event saying that index has been removed. In > addition to that we need to ensure that we don't send any more mgmt > events for such devices and that we don't include such devices in the > response to mgmt_read_index list. > > These are the two places of the code that my patch simplifies, one is > the check for whether to send a "power on" mgmt event and the other the > response handling of read_index_list. I have taken care of that part. It is done inside mgmt_control() and in case of a controller in user channel mode it returns invalid index. For the events, I have no idea on how we would generate events if we are not going through hci_event.c when the user channel is active. Regards Marcel -- To unsubscribe from this list: send the line "unsubscribe linux-bluetooth" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
