Comment 7 for bug 557072

Hi all,

I'm unable to test this machine myself, as I had to return my unit after leaving HP, but I did get some advice on how to make progress on this bug.

The big picture is that the kernel is not putting the hardware into a state that the firmware expects, which is why shutdown/suspend aren't working.

We want to figure out which kernel driver is misbehaving, and to do so, will use the process of elimination.

First step is to simply boot the kernel with init=/bin/bash and verify that shutdown still fails.

Next, repeat step 1, but unload all modules first before shutting down.

Next, see if you can build a kernel with essentially no hardware support, and see if you still get shutdown failures.

Add drivers into your kernel build one-by-one and keep testing shutdown.

Again, the idea here is that some kernel driver is doing the wrong thing, which leaves firmware unable to properly shutdown the machine. By eliminating all the drivers and slowly adding them back one by one, we're trying to figure out which specific driver is misbehaving.

I recommend testing on latest upstream kernel.

Since there are several folks reporting the same bug and you all seem pretty savvy, maybe someone can also try going back multiple versions of the kernel and seeing if one works. If so, that would be a strong hint that there's a broken commit in there, and we can try a bisect too. Please coordinate using this bug or perhaps the mailing list that was setup for members of the Envy 15 launchpad team.

Sorry for all the manual labor involved. This will be tedious, I'm sure. Thanks for all your efforts.