[rt2x00-users] 7 more packets after each beacon frame

[Ivo Van Doorn]

On Mon, Nov 14, 2011 at 3:56 PM, Helmut Schaa
<helmut.schaa at googlemail.com> wrote:
> On Mon, Nov 14, 2011 at 3:51 PM, Ivo Van Doorn <ivdoorn at gmail.com> wrote:
>> On Mon, Nov 14, 2011 at 3:44 PM, Helmut Schaa
>> <helmut.schaa at googlemail.com> wrote:
>>> On Mon, Nov 14, 2011 at 3:38 PM, Ivo Van Doorn <ivdoorn at gmail.com> wrote:
>>>> Hi,
>>>>
>>>>>> Can't we just "program" the card with the right amount of BSSIDs as we
>>>>>> need them?
>>>>>> I.e. just set it to one at the start and change it later when more
>>>>>> BSSIDs are added.
>>>>>
>>>>> Sure, would be possible but if we remove an interface that might leave
>>>>> a hole in the beacon buffers. So, we would have to shift following
>>>>> beacons such that we don't have holes between any beacons ...
>>>>>
>>>>> I'm just thinking about a way to implement this without much management
>>>>> overhead.
>>>>
>>>> Another issue is that this solution will only reduce the number of extra
>>>> beacons. Basically you can program the driver to have:
>>>> 1, 2 or 8 beacons. So when you have 3 interfaces you will still have
>>>> the extra beacons. Furthermore you have to change both the BSSID and
>>>> unicode_to_me mask. And I am not sure if that will show any side effects.
>>>
>>> Hmm, not sure, we could just leave the MAC_BSSID_DW1_BSS_ID_MASK as "8"
>>> and just initialize MAC_BSSID_DW1_BSS_BCN_NUM  with the actual number of
>>> beaconing interfaces. I don't think the BSSID needs to have a specific value
>>> depending on the used beacon slot ...
>>>
>>> So, I think it is possible but quite some changes are required.
>>
>> Yeah, especially since it means you have to constantly ensure that all enabled
>> beacons are placed right after each other in the registers.. I would
>> say that would
>> add a lot of complexity which we don't want. So can we detect which drivers are
>> broken and disable multi-BSSID for those?
>
> I did some more testing with the ath9k_htc monitor and all my devices
> suffer from
> this issue :( I just didn't notice before because my intel monitor
> device just dropped
> all frames with invalid FCS.

Thats bad news :(
But do we still want the multi-bssid for this chipset series?
(or perhaps ask Ralink to fix it in their firmware).

Ivo