Andy Green wrote: > It seems Werner's last patchset (the wake from idle part) has broken PMU > resume reason at the moment due to reading the PMU interrupt registers > we never did before;
Hmm, you should never reach the bulk of that new code (i.e., wait_for_power and friends) on a mere resume. The only thing I changed in the resume path is the UPLL setup. The int1/int2 passing for the power-on reason looks okay as well. (The PMU state would be unreliable for that anyway, since the boot menu test also clears the PMU interrupts. But that's what it always did.) - Werner
