[PATCH 4/5] wcn3620: use new response format for wcn3620 trigger_ba

[Eugene Krasnikov]

Hi Yin,

Yes, it seems like response is FW specific other than chip spesific.

2015-11-12 6:37 GMT+00:00 fengwei.yin <fengwei.yin at linaro.org>:
> Hi Bjorn,
>
> On 2015/11/12 12:50, Bjorn Andersson wrote:
>>
>> On Mon, Nov 9, 2015 at 7:40 AM, Bob Copeland <me at bobcopeland.com> wrote:
>>>
>>> On Mon, Nov 09, 2015 at 05:02:41AM -0500, Yin, Fengwei wrote:
>>>>
>>>> From: Andy Green <andy.green at linaro.org>
>>>>
>>>> From: Andy Green <andy.green at linaro.org>
>>>>
>>>> On wcn3620, firmware response to trigger_ba uses the new, larger
>>>> "v2" format
>>>
>>>
>>>> -     ret = wcn36xx_smd_rsp_status_check(wcn->hal_buf,
>>>> wcn->hal_rsp_len);
>>>> +     ret = wcn36xx_smd_rsp_status_check_v2(wcn, wcn->hal_buf,
>>>> +                                             wcn->hal_rsp_len);
>>>
>>>
>>> It's unclear from the changelog -- is it safe to call
>>> wcn36xx_smd_rsp_status_check_v2 on the 3660/3680 as well?
>>>
>>> Is wcn36xx_smd_rsp_status_check() still needed?
>>>
>>
>> I had to introduce this on one of my 3680 devices recently to silence
>> the error described originally by Andy. So it not only seems safe but
>> seems required. But still, based on how the code was written this
>> doesn't seem to be the case on all versions of the firmware or all
>> chips(?)
>>
> Thanks for the information. It confirm my thought that the change sticks
> to new firmware instead of specific platform.
>
> But we couldn't tell which version of firmware need this new format. Andy's
> original change has two conditions to use the new format:
> 1. The platform is 3620. - this should be removed because you need the same
>    change for 3680. And patch v2 already remove it.
>
> 2. The packet size from firmware is larger than old response size. I suppose
>    this one works in most case.
>
> Regards
> Yin, Fengwei
>
>> Regards,
>> Bjorn
>>
>



-- 
Best regards,
Eugene