Hi Johan,
>>> When we're removing the last item in the white list or adding the first
>>> one to it and HCI_CONNECTABLE is not set we need to update the current
>>> page scan. This patch adds a simple helper function for the purpose and
>>> calls it from the respective mgmt command handlers.
>>>
>>> Signed-off-by: Johan Hedberg <johan.hedberg@xxxxxxxxx>
>>> ---
>>> net/bluetooth/mgmt.c | 31 +++++++++++++++++++++++++++++++
>>> 1 file changed, 31 insertions(+)
>>>
>>> diff --git a/net/bluetooth/mgmt.c b/net/bluetooth/mgmt.c
>>> index 431d5e4aa34b..f5f0763266e0 100644
>>> --- a/net/bluetooth/mgmt.c
>>> +++ b/net/bluetooth/mgmt.c
>>> @@ -5198,6 +5198,24 @@ unlock:
>>> return err;
>>> }
>>>
>>> +/* Helper for Add/Remove Device commands */
>>> +static void update_page_scan(struct hci_dev *hdev, u8 scan)
>>> +{
>>> + if (!test_bit(HCI_BREDR_ENABLED, &hdev->dev_flags))
>>> + return;
>>> +
>>> + if (!hdev_is_powered(hdev))
>>> + return;
>>> +
>>> + if (test_bit(HCI_CONNECTABLE, &hdev->dev_flags))
>>> + return;
>>
>> wouldn't it be better to check the HCI_PSCAN flag here to see if page
>> scan is actually enabled on the controller?
>
> Maybe I was a bit too smart here, but the code is actually correct. We
> also call update_page_scan() with SCAN_DISABLE (0x00) for the
> remove_device case and there if HCI_CONNECTABLE is set we should return
> from the function. Checking for HCI_PSCAN would do the wrong thing.
>
> I can leave the patch as-is, or add a bit more complex logic there:
>
> if (PSCAN && scan)
> return;
>
> if (!PSCAN && !scan)
> return;
>
> Let me know what you prefer.
maybe just add a comment explaining why checking HCI_CONNECTABLE is the right thing to do here.
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
