On 02/11/2021 08:40:53+0900, Dominique Martinet wrote: > Alexandre Belloni wrote on Mon, Nov 01, 2021 at 11:48:46PM +0100: > > On 01/11/2021 10:34:00+0900, Dominique Martinet wrote: > > > Some device cannot be woken up from i2c signal. > > > Add a new irq-gpio attribute for devices which have a gpio connected to > > > the rv8803 INT line so the rtc can be used for suspend to mem > > > > I don't think this is right, the interrupts property of the rtc node can > > point to a gpio and this is expected to be the one connected on INT. You > > don't need another property. > > Oh, why didn't I know about such a useful property. > > I thought I'd have a problem with the device wakeup part but there also > is another 'wakeup-source' property, so there is really nothing left to > do for this patch. > Thank you for the pointer, no code is the best code! > I'd say that you may still need the device_init_wakeup and dev_pm_set_wake_irq calls. -- Alexandre Belloni, co-owner and COO, Bootlin Embedded Linux and Kernel engineering https://bootlin.com
