[RFC] Add the BMPS timer to enter/exit BMPS mode when running
Eugene Krasnikov
k.eugene.e at gmail.com
Wed Jul 24 08:57:56 EDT 2013
> +static void wcn36xx_pmc_bmps_measure_work(struct work_struct *work)
> +{
> + struct wcn36xx *wcn = container_of(work, struct wcn36xx,
> + bmps_work.work);
> +
> + /* Enter or Exit BMPS depend on the currently condition */
> + wcn36xx_pmc_enter_bmps_state(wcn);
> + wcn36xx_pmc_exit_bmps_state(wcn);
> +
> + ieee80211_queue_delayed_work(wcn->hw, &wcn->bmps_work,
> + msecs_to_jiffies(BMPS_MEASURE_TIMER_DEFAULT));
> +}
What is the purpose to queue the work again? Does it mean that wcn36xx
will enter and exit bmps every BMPS_MEASURE_TIMER_DEFAULT seconds?
I also thought you gonna implement a patch that will check if there is
a high data flow then exit bmps . But apparently this patch is not
doing this.
2013/7/24 Olof Johansson <dev at skyshaper.net>:
> Hello Yanbo.
>
> No, I mean this parameter:
> https://www.codeaurora.org/cgit/quic/la/platform/vendor/qcom-opensource/wlan/prima/tree/firmware_bin/WCNSS_qcom_cfg.ini?h=b2g/jb_mr1_rb2.17#n339
>
> Cheers
> --
> Olof
>
> On Wed, Jul 24, 2013 at 4:15 AM, YanBo <dreamfly281 at gmail.com> wrote:
>> Hi Olof,
>>
>> What the specify configuration name you mentioned, do you mean
>> dynamic_dtim? if so, it is used for dynamic adjust the
>> DTIM value but not conflict with BMPS timer
>>
>> BR /Yanbo
>>
>> On Wed, Jul 24, 2013 at 12:59 AM, Olof Johansson <dev at skyshaper.net> wrote:
>>> On Tue, Jul 23, 2013 at 6:46 PM, Eugene Krasnikov <k.eugene.e at gmail.com> wrote:
>>>> @Olof: I am not sure what do you mean under DYNAMIC_PS but there is
>>>> one configuration parameters that is responsible for how time between
>>>> last data packet received from AP and NULL packet sent from STA to AP
>>>> with PS bit set.
>>>
>>> There is a CFG parameter (sent at start or update config smd packets)
>>> which is called DYNAMIC_PS which corresponds to the value of data
>>> inactivity timeout in the prima .ini file. This value adjusts how long
>>> the firmware is staying awake after sending or receiving a packet
>>> before going back to BMPS mode. This value we adjusted several times
>>> to trim throughput as well as btcoex issues.
>>>
>>> At least that is how it was explained to me by QC :)
>>>
>>> Cheers
>>> --
>>> Olof
--
Best regards,
Eugene
More information about the wcn36xx
mailing list