Sven Eckelmann <sven@xxxxxxxxxxxxx> writes: > On Monday, 14 October 2019 00:15:20 CEST Sebastian Gottschall wrote: >> i checked your patch on 10.4 based chipsets with 9984. the values are >> now looking bogus and wrong at all. busy and active time time in ms does >> increase in hours each second >> the problem seem to be that your patch is 10.2.4 only related. >> ath_clean_survey does not trigger on 10.4 so the values double itself >> each time the event raises since you add the full values and not just a >> delta on top > > Thanks for the feedback. So we have now a firmware 10.2.4 which is counting > busy + active up and has wraparound problems. And then we have a 10.4 firmware > (on QCA9888 and QCA4019) which is clearing everything as expected with > WMI_BSS_SURVEY_REQ_TYPE_READ_CLEAR and then we have some 10.4 firmware (one > QCA9984) which behaves more like ath 10.2.4 firmware but is marked as > ATH10K_HW_CC_WRAP_SHIFTED_EACH like the QCA4019. > > So I have no idea how to fix this when QCA4019 and QCA9984 are currently > marked the same but behave differently. Does somebody have a overview how the > different HW versions should behave or is there some special bit in the data > reported by the firmware which can be used to evaluate the expected behavior? I hope there's an easy way to detect this behaviour change, but if nothing else we could add a new bit to enum ath10k_fw_features. But of course that's the last resort, maintaining the firmware features bitfield accross different firmware branches is quite cumbersome. -- https://wireless.wiki.kernel.org/en/developers/documentation/submittingpatches
