Dear Larry, Though my colleague said that they usually use AUTO_REPORT = 1 to run the btcoex. But I think your patch is better because if we just remove the less used part we need to add them back in the future again. So just use your patch but set the AUTO_REPORT default true. Regards, Tony ________________________________________ 從: Larry Finger [larry.finger@xxxxxxxxx] 代表 Larry Finger [Larry.Finger@xxxxxxxxxxxx] 寄件日期: 2017年4月29日 上午 02:52 至: 莊彥宣; Szuyitasi 副本: linux-wireless@xxxxxxxxxxxxxxx; Pkshih; Birming Chiu; Shaofu; Steven Ting 主旨: Re: [PATCH 06/10] rtlwifi: btcoex: 21a 2ant: run mechanism if status changes or auto adjust is set On 04/20/2017 08:30 AM, 莊彥宣 wrote: > Dear SY, > > Linux kernel upstream 上面對於這種 ifdef 的方式是不認同的 > 能否用其他的方式來代替呢? > 或是利用變數來控制。 > > 另外如果default是跑哪邊，某一路只是for special case的話 > 可以先把其中一邊拿掉。 > 先上upstream 之後再來refine > > > Best Regards, > Tony > > -----Original Message----- > From: Kalle Valo [mailto:kvalo@xxxxxxxxxxxxxx] > Sent: Thursday, April 20, 2017 6:54 PM > To: Larry Finger > Cc: linux-wireless@xxxxxxxxxxxxxxx; 莊彥宣; Pkshih; Birming Chiu; Shaofu; Steven Ting > Subject: Re: [PATCH 06/10] rtlwifi: btcoex: 21a 2ant: run mechanism if status changes or auto adjust is set > > Larry Finger <Larry.Finger@xxxxxxxxxxxx> writes: > >> From: Yan-Hsuan Chuang <yhchuang@xxxxxxxxxxx> >> >> The driver will periodically ask the coex, and the coex only runs the >> mechanism when the status was changed or the auto adjust is set. >> >> Signed-off-by: Yan-Hsuan Chuang <yhchuang@xxxxxxxxxxx> >> Signed-off-by: Larry Finger <Larry.Finger@xxxxxxxxxxxx> >> Cc: Pkshih <pkshih@xxxxxxxxxxx> >> Cc: Birming Chiu <birming@xxxxxxxxxxx> >> Cc: Shaofu <shaofu@xxxxxxxxxxx> >> Cc: Steven Ting <steventing@xxxxxxxxxxx> > > [...] > >> @@ -4200,7 +4200,14 @@ void ex_btc8821a2ant_periodical(struct btc_coexist *btcoexist) >> "[BTCoex], ****************************************************************\n"); >> } >> >> +#ifdef BT_AUTO_REPORT_ONLY_8821A_2ANT >> btc8821a2ant_query_bt_info(btcoexist); >> +#else >> btc8821a2ant_monitor_bt_ctr(btcoexist); >> btc8821a2ant_monitor_wifi_ctr(btcoexist); >> + >> + if (btc8821a2ant_is_wifi_status_changed(btcoexist) || >> + coex_dm->auto_tdma_adjust) >> + btc8821a2ant_run_coexist_mechanism(btcoexist); >> +#endif > > This ifdef looks fishy. Why is it needed? > > In general I would expect to see ifdefs only for Kconfig options or defines which value comes from core kernel. > > -- > Kalle Valo Tony, Any resolution of this issue? I would really like to be able to submit the 50 patches I have in my queue. One thought I have is to add boolean variables auto_report_1ant and auto_report_2ant to struct btc_coexist. Then the testing of BT_AUTO_REPORT_ONLY_8821A_1ANT, etc. can be done with an if statement, not with the #if that is currently present. Attached is a patch to fix halbtc8821a1ant. Larry ------Please consider the environment before printing this e-mail.
