On Wednesday 26 April 2006 16:38, Alan Stern wrote: > On Wed, 26 Apr 2006, Rafael J. Wysocki wrote: > > > > Now, if you have specific examples of things that shouldn't be reset, that > > > could be interesting. > > > > The resume device and friends (ie. controller, bus, etc.). > > I presume the freeze/reset we're talking about occurs immediately after > the memory snapshot has been read from the swap partition and immediately > before it is installed/activated. (If occurs before the snapshot is read > in, then how would you actually manage to read it?) > > So under these circumstances, how does it hurt anything to reset the > resume device rather than to freeze it? It just shouldn't be necessary. Actually I think the resume device shouldn't be frozen too. > > My point is that it need not be _necessary_ to reset all devices. We should > > reset only those which need to be reset. > > > Of course every driver should handle this but that's not the point. The point > > is whether we should _force_ the reset on every device. > > I'm not convinced that it hurts anything. But in any case, it wouldn't > hurt to follow Pavel's advice and add a flags field to pm_message_t. Exactly. Greetings, Rafael
