Re: Help me fix a regression caused by 56b9918490 (PM: sleep: Simplify suspend-to-idle control flow)

["Rafael J. Wysocki" <rafael@kernel.org>]

On Fri, Nov 22, 2019 at 1:28 AM Kenneth R. Crudup <kenny@panix.com> wrote:
>
>
> I've been running Linus' tip for a while, mostly to get the PM improvements
> on s2idle (I hate this mode so much- I'm tired of taking a warm laptop out
> of my bag) and I'd reported after a previous merge to Linus' tip from pm-next:
>
> On Wed, 18 Sep 2019, Kenneth R. Crudup wrote:
>
> > - Randomly, if left suspended, nothing other than a hard power off will get
> >   it back  ...  It appears "ec_no_wakeup=1" doesn't have this issue
>
> So I noticed this bug seems to only happen if the laptop is plugged in and
> charging- but only if the battery is < 100%; now that I had a real failure
> mode I started bisecting, and I traced it to 56b991849, "PM: sleep: Simplify
> suspend-to-idle control flow".
>
> What happens is I'll initiate a suspend, then when I try to resume the power
> LED doesn't come on, and the machine is completely unresponsive. I have to
> force a reset by holding the power key.
>
> I'd like to help you fix this issue; I figure if it's affecting my device
> it's affecting others, too. I'm quite proficient with the kernel overall,
> but know very little about ACPI or the EC. I'm assuming- totally wild guess-
> that when we get some of those EC GPEs(?) while suspended(?) we're panic()ing
> or hanging somewhere so the BIOS is getting hung up and the whole system is
> stalled out.
>
> Where should I begin to help debug this? I have EFI store set up so BUG_ON()s
> et al.  will dump there for retrival later if inserting those at critical
> places could help.

First off, please try to change the sleep_no_lps0 ACPI module
parameter (/sys/module/acpi/parameters/sleep_no_lps0) to 1 and see if
that makes any difference.  If it helps, I don't think we can do much
except for blacklisting your machine.