On 17/07/2020 05:27, Tzung-Bi Shih wrote:
I am not convinced the pop comes from 128f825aeab7.
Maybe some pre-existing defect in rk3399_gru_sound got exposed by 128f825aeab7 or the machine driver needs some changes to complement that commit?
(I don't have a rk3399-gru-kevin so I got another test machine with MAX98357A.) (I was testing with and without an audio server.)
Your observations are also a bit different from mine, which IMO also suggests the machine driver is the true culprit -- I'd guess the pops you hear would be from a different problem in your test machine's machine driver?
(Let me restate my observations to contrast with yours, as I feel my previous explanation was too wordy:)
Observations:
- I can hear the pop either with or without 128f825aeab7 (with and
without sdmode-delay).
I never hear pops without 128f825aeab7, but always hear pops with it. (no change when I remove "sdmode-delay" from the device-tree)
- The pop noise is not always. Higher probability after stopping
playback than before starting.
I always hear one pop when starting playback, and two pops (with a few seconds between them) when stopping playback.