On Mon, 2004-11-01 at 18:11 -0700, David Brownell wrote: > On Monday 01 November 2004 13:50, Benjamin Herrenschmidt wrote: > > On Mon, 2004-11-01 at 19:04 +0100, Pavel Machek wrote: > > > Hi! > > > > > > Next try at typechecking for driver layer. I now actually use > > > FREEZE/STANDBY/SUSPEND, altrough I'm not entirely convinced that it is > > > good idea. [There seems to be very little difference between STANDBY > > > and SUSPEND, so it is quite difficult to tell which one should be used > > > when.] > > Erm, "freeze" was never supposed to be a PM state. And I thought > that I took a fair amount of care to write up how "standby" differs > from, for example, "suspend to RAM" ... How so ? it's not a device state, it's a driver state.... but it fits well there too. A driver is "frozen" is a perfectly valid state. A driver is "suspended" implies both frozen and some low power stuff. > > > At this point, we use only suspend I suppose. Standby is just an attempt > > at providing some more distinction for handhelds or such who may have a > > "light" suspend mode with fast wakeup, but it may not make sense ... > > APM talks about "standby" (resembling S1) and "suspend" (resembling S3). > Also "hibernate" (resembling S4). > > Ignoring "standby" states would IMO be a good way to ensure that Linux PM > continues to be inadequate on the hardware that needs PM the most ... :( AM isn't an example here for the reason I already mentioned, the actual low power stuff is under BIOS control. Ben.
