[PATCH] ASoC: rt711-sdca: fix the latency time of clock stop prepare state machine transitions

Shuming [范書銘] shumingf at realtek.com
Thu Nov 17 02:49:55 CET 2022


> > Due to the hardware behavior, it takes some time for CBJ
> detection/impedance sensing/de-bounce.
> > The ClockStop_NotFinished flag will be raised until these functions are
> completed.
> 
> That's fine.
> 
> > In ClockStopMode0 mode case, the SdW controller might check this flag
> > from D3 to D0 when the jack detection interrupt happened.
> 
> That's more confusing. The clock stop sequence is only used in transitions from
> D0 to D3. The timeout is only meant to signal how long the controller driver
> needs to check the device is ready before stopping the clock. The resume
> operation does not require any interaction between controller and device
> drivers.
> 
> Is this an inversion in the wording, or a requirement for the controller driver to
> do something during the resume sequence from D3 to D0?

I wrote it wrongly. It is in the clock stop case, not ClockStopMode0.
This issue was reported by AMD.
If the jack is plugged in/plugged out when SdW controller is in D3 state, the controller will receive the wake event.
They will execute the clock stop exit sequence.



More information about the Alsa-devel mailing list