> > > This looks wrong, you're resyncing the cache and then start resetting > > > registers? It feels like the ordering is off here, and some of this > > > reset sequence might want to be done with the cache bypassed. Are you > > > sure that there's no corruption of user visible state resulting from the > > > power up sequence, especially around the HP driver? > > > Basically the chip runs through the start-up sequence every time in > > the resume function. > > Will move the reset sequence to es8326_suspend > > That's not going to overwrite any user visible settings? Registers written here are internal bias/vmid which should be fixed. User settings like alc and volume won't be overwritten here.
