On Fri, 2021-09-10 at 11:23 +0100, Mark Brown wrote: > On Fri, Sep 10, 2021 at 05:26:13PM +0800, Trevor Wu wrote: > > > When memory allocation for afe->reg_back_up fails, reg_back_up > > can't > > be used. > > Keep the suspend/resume flow but skip register backup when > > afe->reg_back_up is NULL, in case illegal memory access happens. > > It seems like it'd be better to just allocate the buffer at probe > time > and fail in case we can't get it, I'd be surprised if there's many > platforms using this hardware that don't also end up suspending and > resuming. Hi Mark, Thanks for your suggestion. I agree it's better to allocate the memory at probe time. I think we can still keep the implementation in the suspend/resume function as a fallback solution if user doesn't allocate the memory in probe function. In the new mediatek SOCs, regcache has been used to handle register backup. Do I need to add the buffer allocation on probe function to the platform in which mtk_afe_suspend and mtk_afe_resume are used? Thanks, Trevor
