Werner Almesberger werner at openmoko.org
Wed Aug 20 12:57:58 CEST 2008

Andy Green wrote:
> We're talking about OFF state, not suspend.

Yes. What we had happen in GTA01 and HXD8 was power crawling into the
CPU over GPIOs that were still at a high potential. This in turn came
out again on GPIOs going to inputs, such as the PMU's power button
input or the GSM reset/button control transistors, where they caused
further confusion.

So unless we're absolutely sure no GPIO can ever have a non-zero
potential (or only through a very large impedance) while we're in
PMU.STANDBY, or transitioning into or out of it, we shouldn't expect
GPIOs to be in any defined state.

That's why try to control as many signals to subsystems that are
active in PMU.STANDBY as possible from PMU GPIOs. And yes, it would
be great to have more GPIOs that are guaraneteed to have a defined
state in PMU.STANDBY for things like MODEM_RST and MODEM_ON as well.

- Werner

More information about the hardware mailing list