Am Donnerstag 04 Oktober 2007 schrieb Pavel Machek: > On Thu 2007-10-04 12:05:38, Oliver Neukum wrote: > > Am Montag 01 Oktober 2007 schrieb Pavel Machek: > > > > > I guess that the SCSI layer is not really autosuspend-aware, is it? > > > > > > > > Not yet. I would like it to become autosuspend-aware. Oliver wants to > > > > avoid that by keeping all the awareness in usb-storage. (But then what > > > > about all the other SCSI host-adapter drivers? Will they each have > > > > their own way of deciding when a SCSI device can be suspended?) > > > > > > I think I'd prefer to have autosuspending in SCSI layer. That way, we > > > should be able to handle automatic disk spindowns nicely... > > > > OK, so this is the consensus. Then the question arises, how does a > > driver learn that all it's children have gone into autosuspend? > > atomic_t variable in the parent, with children updating it and calling > parent's callback when it goes to 1->0 and 0->1? That's a viable proposal, but it means moving part of the autosuspend code into the generic model. Are you all okay with that? Regards Oliver _______________________________________________ linux-pm mailing list linux-pm@xxxxxxxxxxxxxxxxxxxxxxxxxx https://lists.linux-foundation.org/mailman/listinfo/linux-pm
