On ?t 05-01-06 17:55:33, Preece Scott-PREECE wrote: > I shouldn't oversimplify the power management in a cell phone. When I said we turned whole devices on/off, I was referring only to what the system-level PM (which uses suspend/resume) does. There's a fair amount of subsystem-specific power management outside the Linux suspend/resume framework. Some of it might be handled in the framework, if the framework were more capable. > Ok, system-level PM is not interesting in this discussion. You do runtime-powermanagement of devices, right? Let's take display as an example. (1) Do you have multiple display states like on-fast: can refresh image every 10msec on-slow: can only refresh image once per second, but takes less power ? (2) Do you have multiple display states like off-fast: display is off, user can not see anything, but it only takes msec to wake up off-slow: it takes 1 second for display to wake up, but it takes less power ? [Maybe display is bad example, but I can't think of better device with complex power management] Pavel > scott > > -----Original Message----- > From: Pavel Machek [mailto:pavel@xxxxxx] > Sent: Thursday, January 05, 2006 4:46 PM > To: Preece Scott-PREECE > Cc: Alan Stern; akpm@xxxxxxxx; linux-pm@xxxxxxxxxxxxxx; linux-kernel@xxxxxxxxxxxxxxx > Subject: Re: [linux-pm] [patch] pm: fix runtime powermanagement's /sys interface > > On ?t 05-01-06 17:21:38, Preece Scott-PREECE wrote: > > We do have multiple system-level low-power modes. I believe today they > > differ in turning whole devices on or off, but there are some of those > > devices that could be put in reduced-function/lowered-speed modes if > > we were ready to use a finer-grained distinction. > > I think we were talking multiple off modes for _single device_. It is good to know that even cellphones can get away with whole devices on/off today. > Pavel > > -- > Thanks, Sharp! -- Thanks, Sharp!
