At 6:01 PM -0700 1999/06/17, Jeff Ishaq wrote:
>>12 ms is unacceptable. The entire device should be completely asleep
>>in less than 1ms. In fact, there really isn't much more than 1ms'
>>worth of power available in the supercap.
>
>Prove it! :) Not to sound belligerent at all - I'm asking because I'm
>writing an article on shared libraries, and I'd love to have some cool
>factoids on this issue. Nobody is able to give me actual numbers. I'm not
>much of a hardware guy, or I'd dig them up myself. How big is the supercap?
It varies in size depending on the product, so the library's sleep function must be
coded for worst-case conditions, which is about 1ms. That's only enough time to poke a
couple of hardware registers and return.
>Is there enough information to tell precisely how many instructions we can
>execute when we're running from the supercap?
Nope, because you don't know how big the capacitor is, or how fully it's charged.
>You're right. In my experience, even if you post an event in sleep(), you
>won't see it upon waking. I know the queue is cleared before your sleep()
>is called, but I was never able to see my key even after the device woke
>back up. It just went away.
That was a bug in the pre-alpha ROM you were developing against.
Also, just to clarify, the KEY event queue (NOT the system event queue) is always
flushed when going to sleep, in both "emergency" and "non-emergency" modes.
Regards,
Jim Schram
3Com/Palm Computing
Partner Engineering
