[OLPC-devel] Thoughts on S3 resume.

[Jim Gettys]

On Fri, 2006-08-11 at 18:52 +0100, Chris Ball wrote:

> Looks like we misread this; it's 2.25s, not 0.25s, so would be good to
> know what's happening here.

Yup.  Though .25s amounts are also worth inspection: it is 25 times
longer than the iPAQ takes to resume.  But we'll go for the seconds,
first ;-).

>  
> > around 40 or so to get userspace running again. Doesn't look like the 
> > bottleneck is going to be in the driver code.
> 
> Still need to test the wireless driver, though.

Well, that's a funny one.  Barring USB issues (or something silly in the
TCP stack), we should be able to resume the wireless almost instantly,
as we're normally leaving the wireless powered up running on its own
firmware.  So we clearly won't have to delay for the radio to stabilize
or anything of that order  And we won't have constraints of having a
driver to work on compatible hardware to worry about, since it is a new
driver unique to the chip.

Also, the wakeup on LAN side does *not* depend on USB; we have a
separate signal from the Marvell chip directly to the EC for wakup.
                               - Jim


-- 
Jim Gettys
One Laptop Per Child