On Thursday 28 October 2004 01:26, Patrick Mochel wrote: > > [ Starting to catch up on email.. ] > Me too ... lots of PM-related stuff, it's good to at least have it in one queue now! > I assume that the idea is still to have multiple calls to perform the > entire transition - the first being to queisce the driver and stop any > activity; the second being to perform the actual power transition. > (Actually, there should be three - an additional one in the middle to save > the device states.) > > Do we agree on that? Actually the point I've been meaning to chime in on here is that since, as Benjamin noted, if there's a "write to storage" step (STD, not STR) it uses the "quiesced"/"frozen" system state ... then the real problem is just that the current framework **resumes way too many devices** when writing that out to swap. (Which is a fourth step...) That's relevant here is that it means drivers shouldn't be getting separate "quiesce" then "change power" transitions. That is, there should be no "freeze the driver" request. All drivers except those for the swap device should just go straight into the relevant low power state, and stay. Just like STR. Most of the PITA that I'm seeing with USB should be easily resolved by only resuming the swap device (and ancestors), and that would simplify the suspend model for almost all drivers. (Except for swap devices, and ancestors, which would still need special attentions: appropriate.) - Dave
