[patch] Fix TBTT, TSF handling for TDMA on AR5416 and later chips

[Yin]

Hold on! I just make clean the kernel and rebuilding the kernel.
ath_hal_has_long_rxdesc_tsf() is returning 0 for My DNMA-92 (AR9220) ,
Hence ath_extend_tsf15 is used instead of ath_extend_tsf32 and leads
to beacon stuck. Would you check it please?

ath0: stuck beacon; resetting (bmiss count 4)
ath0: stuck beacon; resetting (bmiss count 4)
ath0: stuck beacon; resetting (bmiss count 4)
ath0: stuck beacon; resetting (bmiss count 4)
ath0: stuck beacon; resetting (bmiss count 4)
ath0: stuck beacon; resetting (bmiss count 4)

Yin

On Wed, Sep 7, 2011 at 11:17 AM, Yin <paradyse at gmail.com> wrote:
> Hi Adrian,
>
> The flag to indicate the capability of 32bit RX TSF is great. It works
> the same as my patch.
>
> Thank you for reviewing it and submitting the patch!
>
> Yin
>
> On Wed, Sep 7, 2011 at 10:38 AM, Yin <paradyse at gmail.com> wrote:
>> Hi Adrian,
>>
>> OK! I will test it now and let you know the result.  :)
>>
>> Thanks,
>> Yin
>>
>> On Wed, Sep 7, 2011 at 9:16 AM, Adrian Chadd <adrian at freebsd.org> wrote:
>>> Hi Yin, all;
>>>
>>> I've done a bit of tidying up of Yin's TDMA related cleanups.
>>>
>>> In particular, I've added a new capability (HAL_CAP_LONG_RXDESC_TSF)
>>> which the AR5416 and later chips set. This indicates the received TSF
>>> is 32 bits rather than 15 bits.
>>>
>>> Yin, would you mind testing this in your local setup and make sure it
>>> works the same as your patch? I'll submit it for inclusion into -HEAD
>>> / 9.0 if it does.
>>>
>>> Thanks again for your great work!
>>>
>>>
>>>
>>> adrian
>>>
>>
>