Hi Amadeusz, Mark > > It is still using snd_soc_{un}register_dai() manually. > > Let's use devm_snd_soc_register_dai(). > > > > Signed-off-by: Kuninori Morimoto <kuninori.morimoto.gx@xxxxxxxxxxx> > > --- (snip) > And while I like the series, after applying all three patches I get > failures in module reload tests. > > If I remember correctly the reason we did explicit flow of register_ and > free_, instead of using devm_register_ is that the lifetime may be > correlated to existence of other device and we need to unregister at > correct moment. Maybe something to revisit, but I would prefer to keep > tests green for now. I have sent patch-set to use devm_xxx as much as possible before, but there are still some drivers which doesn't use it. It is because of lifetime, like this time. I don't remember detail and/or which driver was, and no comment about it on code. But this 3 patches might be it. Let's keep all tests green. Please drop this patch-set > Mark Sorry for my noise Thank you for your help !! Best regards --- Kuninori Morimoto