Comment 9 for bug 904261

So, this basically means that the S3 test script observed that the kernel wrote to the power management registers which then transitions the machine into the S3 state.

The kernel doesn't do much more after this it is waiting for the machine to suspend. So a hang like this is quite curious. If S5 (shutdown) works, then this could indicate that the PM1 controls don't work but as discussed on IRC, S5 does work, so it is strange that transitioning to S3 breaks.

I'm out of ideas. We should pull in Alex Hung as he may have some notion of what to do next.