On Wed, May 16, 2018 at 10:29:38AM +0200, Rafael J. Wysocki wrote: > On Wed, May 16, 2018 at 12:18 AM, Dmitry Torokhov > <dmitry.torokhov@xxxxxxxxx> wrote: > > On Wed, Feb 28, 2018 at 3:09 AM, Rafael J. Wysocki <rjw@xxxxxxxxxxxxx> wrote: > >> > >> From: Rafael J. Wysocki <rafael.j.wysocki@xxxxxxxxx> > >> > >> The intel-vbtn device will not be able to wake up the system any more > >> after removing the notify handler provided by its driver, so make > >> its sysfs attributes reflect that. > > > > Are there devices that can and should wake up system even after driver > > is unbound from device? > > Maybe, but then something needs to configure them for wakeup on suspend, right? Yes. I don't know of any such instances though, that's why I asked. > > > Should we move resetting wakeup flag into the > > driver core so we do not have to do it in each individual driver? > > It could be done in the driver core or bus type. The bus type level > is a better place probably, because of PCI. > > Anyway, the idea of this patch was to clear the bit in the same entity > that set it. :-) Yeah, I understand. I was just thinking that with devm* API we are able to remove most if not all remove() code in drivers, and similarly to resetting driver data pointers, it would be great if we moved resetting of wakeup flags into either bus implementation or device core itself. Thanks. -- Dmitry
