RE: [PATCH rtw-next 04/12] wifi: rtw89: mac: reset power state before switching to power on

[Ping-Ke Shih <pkshih@realtek.com>]

Bitterblue Smith <rtl8821cerfe2@gmail.com> wrote;
> On 12/12/2025 05:12, Ping-Ke Shih wrote:
> > To run power on function properly, reset power states (off/on/PS) to
> > initial state. Otherwise, it might be unusable due to fail to power on.
> >
> > Signed-off-by: Ping-Ke Shih <pkshih@realtek.com>
> > ---
> >  drivers/net/wireless/realtek/rtw89/mac.c    |  89 ++++++++++++--
> >  drivers/net/wireless/realtek/rtw89/mac.h    |   1 +
> >  drivers/net/wireless/realtek/rtw89/mac_be.c | 130 ++++++++++++++++++++
> >  drivers/net/wireless/realtek/rtw89/reg.h    |  27 ++++
> >  4 files changed, 240 insertions(+), 7 deletions(-)
> >
> > diff --git a/drivers/net/wireless/realtek/rtw89/mac.c b/drivers/net/wireless/realtek/rtw89/mac.c
> > index d78fbe73e365..ba4344a640f5 100644
> > --- a/drivers/net/wireless/realtek/rtw89/mac.c
> > +++ b/drivers/net/wireless/realtek/rtw89/mac.c
> > @@ -1478,13 +1478,11 @@ static void rtw89_mac_power_switch_boot_mode(struct rtw89_dev *rtwdev)
> >
> >  static int rtw89_mac_power_switch(struct rtw89_dev *rtwdev, bool on)
> >  {
> > -#define PWR_ACT 1
> >       const struct rtw89_mac_gen_def *mac = rtwdev->chip->mac_def;
> >       const struct rtw89_chip_info *chip = rtwdev->chip;
> >       const struct rtw89_pwr_cfg * const *cfg_seq;
> >       int (*cfg_func)(struct rtw89_dev *rtwdev);
> >       int ret;
> > -     u8 val;
> >
> >       rtw89_mac_power_switch_boot_mode(rtwdev);
> >
> > @@ -1499,10 +1497,10 @@ static int rtw89_mac_power_switch(struct rtw89_dev *rtwdev, bool on)
> >       if (test_bit(RTW89_FLAG_FW_RDY, rtwdev->flags))
> >               __rtw89_leave_ps_mode(rtwdev);
> >
> > -     val = rtw89_read32_mask(rtwdev, R_AX_IC_PWR_STATE, B_AX_WLMAC_PWR_STE_MASK);
> > -     if (on && val == PWR_ACT) {
> > -             rtw89_err(rtwdev, "MAC has already powered on\n");
> > -             return -EBUSY;
> > +     if (on) {
> > +             ret = mac->reset_pwr_state(rtwdev);
> > +             if (ret)
> > +                     return ret;
> >       }
> >
> >       ret = cfg_func ? cfg_func(rtwdev) : rtw89_mac_pwr_seq(rtwdev, cfg_seq);
> > @@ -1529,7 +1527,6 @@ static int rtw89_mac_power_switch(struct rtw89_dev *rtwdev, bool on)
> >       }
> >
> >       return 0;
> > -#undef PWR_ACT
> >  }
> >
> >  int rtw89_mac_pwr_on(struct rtw89_dev *rtwdev)
> > @@ -3931,6 +3928,83 @@ static int rtw89_mac_feat_init(struct rtw89_dev *rtwdev)
> >       return 0;
> >  }
> >
> > +static int rtw89_mac_reset_pwr_state_ax(struct rtw89_dev *rtwdev)
> > +{
> > +     enum rtw89_core_chip_id chip_id = rtwdev->chip->chip_id;
> > +     u32 val32;
> > +     int ret;
> > +
> > +     val32 = rtw89_read32(rtwdev, R_AX_SYSON_FSM_MON);
> > +     val32 &= WLAN_FSM_MASK;
> > +     val32 |= WLAN_FSM_SET;
> > +     rtw89_write32(rtwdev, R_AX_SYSON_FSM_MON, val32);
> > +
> > +     ret = read_poll_timeout(rtw89_read32_mask, val32, val32 == WLAN_FSM_IDLE,
> > +                             1000, 2000000, false,
> > +                             rtwdev, R_AX_SYSON_FSM_MON, WLAN_FSM_STATE_MASK);
> > +     if (ret) {
> > +             rtw89_err(rtwdev, "[ERR]Polling WLAN PMC timeout= %X\n", val32);
> > +             return ret;
> > +     }
> 
> I get this error message with every AX chip the first time
> rtw89_mac_pwr_on() is called:
> 
> [  +0.007286] rtw89_8852bu 2-4:1.0: loaded firmware rtw89/rtw8852b_fw-1.bin
> [  +2.001260] rtw89_8852bu 2-4:1.0: [ERR]Polling WLAN PMC timeout= 100
> 
> They still work, because the second attempt in rtw89_mac_pwr_on()
> succeeds.
> 
> Before this patch, the message was "MAC has already powered on".
> The new message says it's an error, so it will confuse people.

I borrow a 8852AU and a 8852CU that both don't have error messages.
The "first time" you meant is just to plug the USB adapter?
If so, I have not idea about the difference between yours and mine.

> 
> Is it possible this code related to R_AX_SYSON_FSM_MON is not
> applicable to USB? My adapters still work if I delete it.

I'm checking internal experts. Quickly checking vendor driver, I don't see
it avoids calling this because of USB devices. 

If you delete code related to R_AX_SYSON_FSM_MON, will it show below messages
under the case 'val32 == MAC_AX_MAC_ON'?

   "[ERR]Polling MAC state timeout= %X\n"

> 
> (I don't get the error with RTL8912AU.)
> 

I'm adding this patch for WiFi 7 chips mainly, and this seems to be expected.