Mini-Conference Proposal: Suspend/Resume speed

[Richard A. Smith]

I propose that for Update.2 we put significant effort into making 
suspend/resume reach its speed target and work correctly on the hardware.

Its close but we are still about 5x slower than the 200ms goal, things 
like SD don't always work after a cycle, and we need to have proper idle 
detection.

This is not to be confused with all the incompatibilities between 
presence and suspend or network apps that don't work when a peer has 
suspended.  Those problems are in a different space.

Those problems certainly need to be solved but trying to solve them on 
top of a suspend/resume implementation that not the final product just 
makes it harder.

-- 
Richard Smith  <richard at laptop.org>
One Laptop Per Child