Hi. Pavel Machek wrote: > On Wed 2010-01-27 01:51:03, Rafael J. Wysocki wrote: >> On Tuesday 26 January 2010, Pavel Machek wrote: >>>>>>> The ideal behavior would be: >>>>>>> >>>>>>> if(disk is spun up) >>>>>>> then let the sync happen >>>>>> I'm not against that. Patch welcome. :-) >>>>> I'd say such knob would be ugly. >>>> Define "ugly", please. >>> Per-system property, which should better be >>> per-program-that-requires-suspend. You request suspend without syncing >>> (you want it quick, battery is 90%), then the battery runs low, and >>> system daeomn requests s2ram, not realizing that someone disabled sync >>> from under him. >> I really prefer a per-system setting. The program that wants to sync anyway >> can easily do that by itself. > > Well, existing programs expect existing behaviour... Programs that do > not want to sync can easily do it themselves, too, without afecting > rest of system. What's all this talk about 'programs'? The direct invokers of the hibernation and suspend-to-ram code are few in number. They're easily modified to match changes in the kernel code, and when it comes to syncing, there's no need for the modification to be made in lock-step. If you hibernate or suspend without having dirty pages synced to disk first, it shouldn't matter, should it? Resuming is pretty reliable, and even if it fails, journal recovery is not problematic. Regards, Nigel _______________________________________________ linux-pm mailing list linux-pm@xxxxxxxxxxxxxxxxxxxxxxxxxx https://lists.linux-foundation.org/mailman/listinfo/linux-pm
