> > > > > > + cancel_delayed_work_sync(&sma1303->check_fault_work); > > > > > > > + msleep(55); > > > > > > That sleep looks odd - what are we delaying after? > > > > > It need for IC(Amp) issue. > > > > Right, but what is the issue? It's not clear what event we're > > > delaying for so it's not clear it'll work properly. > > > This device requires a delay from mute to power off. If it is turned > > off while the signal is coming out, it takes some time to become > > muted. (need to about 50ms). So, this delay is required to turn off > > the power after completely muting. If the time is too long, we can > > reduce the delay a bit. > > This at the very least needs a comment since it's not clear how someone > would work that out from the code, there's no connection here to the mute > operation so someone might just remove the delay. It may even be that this > is happening well over 55ms after the mute happened. The ideal thing would > be to track the time from the last mute, the regulator core does something > along those lines with off_on_delay. Currently mute control is registered in mute_stream function in dai operations. After that, amp shutdown is performed in DAPM PRE_PMD. Since a 55 ms delay was required in the meantime, it seems correct to change the position of the delay like you said. I will also include a comment about delay. Please reply if there is any problem. > Please fix your mail client to word wrap within paragraphs at something > substantially less than 80 columns. Doing this makes your messages much > easier to read and reply to. Thanks, I fixed the mail setting.
Attachment:
signature.asc
Description: signature.asc