Hi Tobias, Am Freitag, 28. Februar 2020, 15:57:10 CET schrieb Tobias Schramm: > thanks for the review. I'll implement the changes and send a v2. > > >> + * of wakeup sources without disabling the whole key > > Also can you explain the problem a bit? If there is a deficit in the input > > subsystem regarding wakeup events, dt is normally not the place to work > > around things [we're supposed to be OS independent] > > The issue is that some users wanted to be able to control the wakeup > functionality of the keys separately via sysfs. That does not seem to be > possible when combining both keys into one gpio-keys node. A more > detailed explanation of the issue can be found at [1]. ok ... but that is really strange, because looking at gpio-keys.c I see it checking the individual button wakeup-property before setting the irq-wake in gpio_keys_enable_wakeup() . Ah, but I guess manually disabling/enabling wakeup via sysfs only works for the whole device and all wakeup buttons. In general this sounds more like a gpio-keys deficit, but in the end we can keep the separate gpio-key nodes here, they don't violate any dt-bindings ;-) . Heiko
