On Thu, 3 Dec 2020 at 02:29, Nicolas Boichat <drinkcat@xxxxxxxxxxxx> wrote: > > This causes a 0-day warning (on our chromeos-5.4 backports but I don't > see why upstream would not be affected): > https://groups.google.com/g/cros-kernel-buildreports/c/MfS3SInT5jg/m/Hkzxh_U7AwAJ > > Didn't look at the details of the config, but SET_SYSTEM_SLEEP_PM_OPS > is a noop if CONFIG_PM_SLEEP is not set (while SET_RUNTIME_PM_OPS is > noop-ed by CONFIG_PM). > > So I guess msdc_suspend/msdc_resume should be guarded by > CONFIG_PM_SLEEP instead of CONFIG_PM. Yep. There is already a patch [1] in my queue for this. It uses the _maybe_unused approach (Arnd prefers that and I don't mind). [...] Kind regards Uffe [1] https://patchwork.kernel.org/project/linux-mediatek/patch/20201203222922.1067522-1-arnd@xxxxxxxxxx/
